Revert the revert of 275818: "dcommit binaries for test for code review 310483003"

Revert the revert of 275818: "dcommit binaries for test for code review 310483003"

Reverts: https://codereview.chromium.org/323923002/
Which reverted: https://codereview.chromium.org/310483003/

The revert was necessary because the .so files in this patch apparently need
to have their executable bits set, even though they are just test data that
isn't actually executable. For posterity, the presubmit scripts were run with
"-v" locally, but did not show the error that the bots eventually produced.

BUG=
R=simonb@google.com

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=275833

[Andrew Hayden (chromium.org), 2014-06-09 17:30:24]

Sorry about that. This time, presubmit scripts are apparently happy (last time,
they were complaining on my machine about missing OWNERS, which wasn't an issue
for this checkin; I don't know why that masked the executable bit stuff, but it
did).

This time it looks sane: the "new file mode" stuff is in the diff, so I assume
SVN will be happy as well:

src/tools/relocation_packer/test_data$ git diff HEAD^ HEAD
diff --git a/tools/relocation_packer/test_data/elf_file_unittest_relocs.so
b/tools/relocation_packer/test_data/elf_file_unittest_relocs.so
new file mode 100755
index 0000000..dc39cfc
Binary files /dev/null and
b/tools/relocation_packer/test_data/elf_file_unittest_relocs.so differ
diff --git
a/tools/relocation_packer/test_data/elf_file_unittest_relocs_packed.so
b/tools/relocation_packer/test_data/elf_file_unittest_relocs_packed.so
new file mode 100755
index 0000000..a0f064e

[simonb1, 2014-06-09 17:32:35]

lgtm

[Andrew Hayden (chromium.org), 2014-06-09 17:33:08]

Committed patchset #1 manually as r275833 (presubmit successful).