AU: Payload Signer class

AU: Payload Signer class

This class can take a private key and sign a blob of data. The API is
amenable to the upcoming change to delta_diff_generator that will use
it.

Also, minor change to the protobuf to support signatures.

TEST=unittests
BUG=5662

[petkov, 2010-08-20 05:28:05]

Looks OK functionally, some nits and suggestions below.

LGTM

http://codereview.chromium.org/3173032/diff/1/3
File payload_signer.cc (right):

http://codereview.chromium.org/3173032/diff/1/3#newcode1
payload_signer.cc:1: // Copyright (c) 2010 The Chromium Authors. All rights
reserved.
Chromium OS Authors

http://codereview.chromium.org/3173032/diff/1/3#newcode6
payload_signer.cc:6: #include "base/logging.h"
insert an empty line before base/logging.h

http://codereview.chromium.org/3173032/diff/1/3#newcode30
payload_signer.cc:30: vector<string> cmd;
It might be better to use base::SplitString (from string_util.h) here to split a
command line string into vector<string>. Then you wouldn't need the comment
above about the actual command line.

http://codereview.chromium.org/3173032/diff/1/4
File payload_signer.h (right):

http://codereview.chromium.org/3173032/diff/1/4#newcode1
payload_signer.h:1: // Copyright (c) 2010 The Chromium Authors. All rights
reserved.
Chromium OS Authors

http://codereview.chromium.org/3173032/diff/1/4#newcode14
payload_signer.h:14: // blob, which should be appended. See
update_metadata.proto for more info.
... and returns the signature blob, which should be appended...

http://codereview.chromium.org/3173032/diff/1/5
File payload_signer_unittest.cc (right):

http://codereview.chromium.org/3173032/diff/1/5#newcode19
payload_signer_unittest.cc:19: const char* kUnittestPrivateKeyPath =
"unittest_key.pem";
const char kUnit...[] = ...

http://codereview.chromium.org/3173032/diff/1/5#newcode25
payload_signer_unittest.cc:25: TEST(PayloadSignerTest, SimpleTest) {
Given that you've defined the fixture class PayloadSignerTest, you should use
TEST_F, I think. Otherwise, you don't need to declare the PayloadSignerTest... I
may be wrong.

http://codereview.chromium.org/3173032/diff/1/5#newcode26
payload_signer_unittest.cc:26: // static bool SignPayload(const std::string&
unsigned_payload_path,
Do you really need these comments here? Remove, I'd say...

http://codereview.chromium.org/3173032/diff/1/5#newcode56
payload_signer_unittest.cc:56: 
Remove blank line.

http://codereview.chromium.org/3173032/diff/1/5#newcode62
payload_signer_unittest.cc:62: kDataToSign.c_str(),
kDataToSign.data()

http://codereview.chromium.org/3173032/diff/1/5#newcode77
payload_signer_unittest.cc:77:
EXPECT_TRUE(signatures.ParseFromArray(&signature_blob[0],
signature_blob.data() maybe?

http://codereview.chromium.org/3173032/diff/1/5#newcode83
payload_signer_unittest.cc:83: EXPECT_EQ(sizeof(kExpectedSignature),
sig_data.size());
Can't you switch to ASSERT_EQ and avoid the if below?