[test262] Use standalone Test262 harness project

test/test262/test262.gyp

Index: test/test262/test262.gyp
diff --git a/test/test262/test262.gyp b/test/test262/test262.gyp
index 2bdc6cefe5092c5265caadb0f2acecceb078ab21..605a5ee7cbc039e3004f0eb8fa52dae24a383291 100644
--- a/test/test262/test262.gyp
+++ b/test/test262/test262.gyp
@@ -23,7 +23,7 @@
             {
               'action_name': 'archive_test262',
               'inputs': ['archive.py', '<!@(python list.py)'],
-              'outputs': ['data.tar'],
+              'outputs': ['data.tar', 'harness.tar'],

[tandrii(chromium), 2016/07/20 10:42:01]

Michael has to approve this, bcz:

1. this does increase the size of things to be downloaded on bot,   do you know
by how much? Is untarring done always, or only on bots that run harness?
2. Is there any reason to ship it separately in harness.tar and not update
existing data.tar?
3. You probably have to prepare test set (~ a bunch of isolate hashes) for
Michael's end-to-end test.

[Michael Achenbach, 2016/07/20 11:22:49]

1. Insignificant. As described, it was part of the test262 data before. It
probably is right now until test262 is rolled.
2. +1, should just be packed into data tar IMO.
3. Test262 is not part of the end-to-end test -> ignore.

               'action': ['python', 'archive.py'],
             },
           ],