Revert "Roll src/third_party/boringssl/src 68de407:65a7e94"

Revert "Roll src/third_party/boringssl/src 68de407:65a7e94"

This reverts https://codereview.chromium.org/1122423007 , it broke
the android clang and the official mac bots:
http://build.chromium.org/p/chromium.linux/builders/Android%20Clang%20Builder%20%28dbg%29/builds/59288
http://build.chromium.org/p/chromium.chrome/builders/Google%20Chrome%20Mac/builds/1025

BUG=none

[Nico, 2015-05-09 04:07:06]

thakis@chromium.org changed reviewers:
+ davidben@chromium.org

[Nico, 2015-05-09 04:07:07]

Not sure why your revert isn't in yet. Maybe I'll find out in a second…

[Nico, 2015-05-09 04:08:26]

Jinx. Nevermind then, thanks for the revert :-)

[davidben, 2015-05-09 04:17:12]

You probably wouldn't have been able to land this one anyway. I had to re-upload
it with this lovely patch to depot_tools:

diff --git a/third_party/upload.py b/third_party/upload.py
index eac2a3d..de6a9fb 100755
--- a/third_party/upload.py
+++ b/third_party/upload.py
@@ -1062,7 +1062,7 @@ class VersionControlSystem(object):
     """Returns true if data contains a null byte."""
     # Derived from how Mercurial's heuristic, see
     # http://selenic.com/hg/file/848a6658069e/mercurial/util.py#l229
-    return bool(data and "\0" in data)
+    return bool(data and ("\0" in data) or "Karatsuba post-processing" in data)
 
 
 class SubversionVCS(VersionControlSystem):

That's part of why it took a bit longer for my revert to get through. Some of
those files have Latin-1 characters that make Rietveld barf because thank you
OpenSSL and Python 2 strings. :-) (I'll probably try to get rid of them later.
But they're already in the repository because Rietveld *doesn't* break on new
files, only diffs, and we didn't notice until it caused diff problems.)