CT: First step towards supporting Certificate Transparency in Chrome.

net/cert/signed_certificate_timestamp.h

Index: net/cert/signed_certificate_timestamp.h
diff --git a/net/cert/signed_certificate_timestamp.h b/net/cert/signed_certificate_timestamp.h
new file mode 100644
index 0000000000000000000000000000000000000000..5165f20f19ed1a46d9c084783e32ec42fe9e4d62
--- /dev/null
+++ b/net/cert/signed_certificate_timestamp.h
@@ -0,0 +1,98 @@
+// Copyright (c) 2013 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef NET_CERT_SIGNED_CERTIFICATE_TIMESTAMP_H_
+#define NET_CERT_SIGNED_CERTIFICATE_TIMESTAMP_H_
+
+#include <string>
+#include <vector>
+
+#include "base/time/time.h"
+#include "net/base/hash_value.h"
+#include "net/base/net_export.h"
+
+namespace net {
+
+// Structures related to Certificate Transparency (RFC6962).
+namespace ct {
+
+struct NET_EXPORT LogEntry {

[wtc, 2013/10/24 23:14:23]

General suggestion: to help future maintainers of this code, I suggest you
include the definitions of the original types in comments. For example, for
struct LogEntry, add:

       enum { x509_entry(0), precert_entry(1), (65535) } LogEntryType;

       struct {
           LogEntryType entry_type;
           select (entry_type) {
               case x509_entry: X509ChainEntry;
               case precert_entry: PrecertChainEntry;
           } entry;
       } LogEntry;

If you think this is too verbose, at least mention the name of the types (for
exmple, "enum LogEntryType and struct LogEntry in RFC 6962, Sec. 3.1") in
comments.

[Eran M. (Google), 2013/10/30 18:00:08]

Resolved by mentioning the name of RFC & section for each of the relevant
structs/enums.

+  enum Type {
+    LOG_ENTRY_TYPE_X509 = 0,
+    LOG_ENTRY_TYPE_PRECERT = 1
+  };
+
+  LogEntry();
+  ~LogEntry();
+
+  void Reset();
+
+  Type type;

[wtc, 2013/10/24 23:14:23]

You should use a union below.

[Eran M. (Google), 2013/10/30 18:00:08]

the objects in the suggested union are not plain data types and so are forbidden
(unrestricted unions are available with c++0x only).

I could use pointers or boost::variant, if you think it's important enough.

+
+  // Set if type == LOG_ENTRY_TYPE_X509
+  std::string leaf_cert;

[wtc, 2013/10/24 23:14:23]

Nit: leaf_cert => leaf_certificate

because you didn't abbreviate "tbs_certificate".

[Eran M. (Google), 2013/10/30 18:00:08]

Done.

+
+  // Set if type == LOG_ENTRY_TYPE_PRECERT
+  SHA256HashValue issuer_key_hash;
+  std::string tbs_certificate;
+};
+
+// Helper structure to represent Digitally Signed data, as described in
+// Sections 4.7 and 7.4.1.4.1 of RFC 5246.
+struct DigitallySigned {
+  enum HashAlgorithm {
+    HASH_ALGO_NONE = 0,
+    HASH_ALGO_MD5 = 1,
+    HASH_ALGO_SHA1 = 2,
+    HASH_ALGO_SHA224 = 3,
+    HASH_ALGO_SHA256 = 4,
+    HASH_ALGO_SHA384 = 5,
+    HASH_ALGO_SHA512 = 6,
+  };
+
+  enum SignatureAlgorithm {
+    SIG_ALGO_ANONYMOUS = 0,
+    SIG_ALGO_RSA = 1,
+    SIG_ALGO_DSA = 2,
+    SIG_ALGO_ECDSA = 3
+  };
+
+  DigitallySigned();
+  ~DigitallySigned();
+
+  HashAlgorithm hash_algorithm;
+  SignatureAlgorithm signature_algorithm;
+  std::string signature_data;

[wtc, 2013/10/24 23:14:23]

Nit: this field is named "signature" in the RFC. If you deliberately renamed it
"signature_data" for clarity, that's fine by me.

[Eran M. (Google), 2013/10/30 18:00:08]

I have (found it confusing when actually accessing the field).

+};
+
+struct NET_EXPORT SignedCertificateTimestamp {
+  enum Version {
+    SCT_VERSION_1 = 0,
+  };
+
+  // Source of the SCT - supplamentary, not defined in CT RFC.

[wtc, 2013/10/24 23:14:23]

Typo: supplamentary => supplementary

[Eran M. (Google), 2013/10/30 18:00:08]

Done.

+  enum Origin {
+    EMBEDDED = 0,
+    FROM_TLS_HANDSHAKE = 1,
+    FROM_OCSP_RESPONSE = 2,

[wtc, 2013/10/24 23:14:23]

Nit: the naming of the Origin enumerators differs in style from the other
enumerators in this header, which all have a consistent prefix that suggests
which enum type they belong to. Either style is fine by me. I am just pointing
out this inconsistency.

[Eran M. (Google), 2013/10/30 18:00:08]

Done - changed to conform other enums.

+  };
+
+  SignedCertificateTimestamp();
+  ~SignedCertificateTimestamp();
+
+  Version version;
+  std::string log_id;
+  base::Time timestamp;
+  std::string extensions;
+  DigitallySigned signature;
+  // The origin should not participate in equality checks
+  // As the same SCT can be provided from multiple sources.

[wtc, 2013/10/24 23:14:23]

Nit: As => as

[Eran M. (Google), 2013/10/30 18:00:08]

Done.

+  Origin origin;
+};
+
+}  // namespace ct
+
+}  // namespace net
+
+#endif  // NET_CERT_SIGNED_CERTIFICATE_TIMESTAMP_H_