Improve testing for SDCH.

sdch/sdch.gyp

Index: sdch/sdch.gyp
diff --git a/sdch/sdch.gyp b/sdch/sdch.gyp
index 233ac534d2f7124f6c39bc359a9f1204264fa3f6..52a92f31e2a7a55c535d0ad313e60a585dca8b1b 100644
--- a/sdch/sdch.gyp
+++ b/sdch/sdch.gyp
@@ -29,15 +29,19 @@
         'open-vcdiff/src/encodetable.h',
         'open-vcdiff/src/google/output_string.h',
         'open-vcdiff/src/google/vcdecoder.h',
+        'open-vcdiff/src/google/vcencoder.h',
         'open-vcdiff/src/headerparser.cc',
         'open-vcdiff/src/headerparser.h',
         'open-vcdiff/src/instruction_map.cc',
         'open-vcdiff/src/instruction_map.h',
+        'open-vcdiff/src/jsonwriter.h',

[jar (doing other things), 2014/07/31 23:19:23]

I didn't think to read this carefully.... but since you asked about my thought
on the "sdch" directory... and this is the only such entry...

Why are we using their jsonwriter?  Don't we have such functionality in
Chromium?  

take a look in base/json/*

base/json/json_writer.cc for example.

[Randy Smith (Not in Mondays), 2014/08/11 20:33:25]

To be explicit, the things I need sdch/OWNERS review on are this change and the
inclusion of it in chrome/browser/net/DEPS.
The issue is that, in adding the vcencoder.cc file, I ended up with a undefined
symbol which was to the jsonwriter.  I'm very reluctant to try and hook the
vcencoder.cc file up to our jsonwriter if it was written in parallel with and to
depend on the SDCH jsonwriter file.

+        'open-vcdiff/src/jsonwriter.cc',
         'open-vcdiff/src/rolling_hash.h',
         'open-vcdiff/src/testing.h',
         'open-vcdiff/src/varint_bigendian.cc',
         'open-vcdiff/src/varint_bigendian.h',
         'open-vcdiff/src/vcdecoder.cc',
+        'open-vcdiff/src/vcencoder.cc',
         'open-vcdiff/src/vcdiff_defs.h',
         'open-vcdiff/src/vcdiffengine.cc',
         'open-vcdiff/src/vcdiffengine.h',