Move email and domain canonicalization helper to common/net/gaia

This is in preparation for a CL that uses this functionality in non-chromeos code.

BUG=None
TEST=Compiles and passes tests, no functional changes.
R=nkostylev@chromium.org
TBR=arv@chromium.org


Review URL: https://chromiumcodereview.appspot.com/10584039

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@143899 0039d316-1c4b-4281-b951-d872f2087c98

3 files changed, 180 insertions, 0 deletions

diff --git a/chrome/common/net/gaia/gaia_auth_util.cc b/chrome/common/net/gaia/gaia_auth_util.cc
new file mode 100644
index 0000000..ec9f421
--- /dev/null
+++ b/chrome/common/net/gaia/gaia_auth_util.cc
@@ -0,0 +1,61 @@
+// Copyright (c) 2012 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.
+
+#include "chrome/common/net/gaia/gaia_auth_util.h"
+
+#include <vector>
+
+#include "base/logging.h"
+#include "base/string_split.h"
+#include "base/string_util.h"
+
+namespace gaia {
+
+namespace {
+const char kGmailDomain[] = "gmail.com";
+}
+
+std::string CanonicalizeEmail(const std::string& email_address) {
+  std::vector<std::string> parts;
+  char at = '@';
+  base::SplitString(email_address, at, &parts);
+  if (parts.size() != 2U)
+    NOTREACHED() << "expecting exactly one @, but got " << parts.size();
+  else if (parts[1] == kGmailDomain)  // only strip '.' for gmail accounts.
+    RemoveChars(parts[0], ".", &parts[0]);
+  std::string new_email = StringToLowerASCII(JoinString(parts, at));
+  VLOG(1) << "Canonicalized " << email_address << " to " << new_email;
+  return new_email;
+}
+
+std::string CanonicalizeDomain(const std::string& domain) {
+  // Canonicalization of domain names means lower-casing them. Make sure to
+  // update this function in sync with Canonicalize if this ever changes.
+  return StringToLowerASCII(domain);
+}
+
+std::string SanitizeEmail(const std::string& email_address) {
+  std::string sanitized(email_address);
+
+  // Apply a default domain if necessary.
+  if (sanitized.find('@') == std::string::npos) {
+    sanitized += '@';
+    sanitized += kGmailDomain;
+  }
+
+  return sanitized;
+}
+
+std::string ExtractDomainName(const std::string& email_address) {
+  // First canonicalize which will also verify we have proper domain part.
+  std::string email = CanonicalizeEmail(email_address);
+  size_t separator_pos = email.find('@');
+  if (separator_pos != email.npos && separator_pos < email.length() - 1)
+    return email.substr(separator_pos + 1);
+  else
+    NOTREACHED() << "Not a proper email address: " << email;
+  return std::string();
+}
+
+}  // namespace gaia
diff --git a/chrome/common/net/gaia/gaia_auth_util.h b/chrome/common/net/gaia/gaia_auth_util.h
new file mode 100644
index 0000000..2d4ab00
--- /dev/null
+++ b/chrome/common/net/gaia/gaia_auth_util.h
@@ -0,0 +1,32 @@
+// Copyright (c) 2012 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 CHROME_COMMON_NET_GAIA_GAIA_AUTH_UTIL_H_
+#define CHROME_COMMON_NET_GAIA_GAIA_AUTH_UTIL_H_
+#pragma once
+
+#include <string>
+
+namespace gaia {
+
+// Perform basic canonicalization of |email_address|, taking into account that
+// gmail does not consider '.' or caps inside a username to matter. It also
+// ignores everything after a '+'. For example, c.masone+abc@gmail.com ==
+// cMaSone@gmail.com, per
+// http://mail.google.com/support/bin/answer.py?hl=en&ctx=mail&answer=10313#
+std::string CanonicalizeEmail(const std::string& email_address);
+
+// Returns the canonical form of the given domain.
+std::string CanonicalizeDomain(const std::string& domain);
+
+// Sanitize emails. Currently, it only ensures all emails have a domain by
+// adding gmail.com if no domain is present.
+std::string SanitizeEmail(const std::string& email_address);
+
+// Extract the domain part from the canonical form of the given email.
+std::string ExtractDomainName(const std::string& email);
+
+}  // namespace gaia
+
+#endif  // CHROME_COMMON_NET_GAIA_GAIA_AUTH_UTIL_H_
diff --git a/chrome/common/net/gaia/gaia_auth_util_unittest.cc b/chrome/common/net/gaia/gaia_auth_util_unittest.cc
new file mode 100644
index 0000000..0f7c702
--- /dev/null
+++ b/chrome/common/net/gaia/gaia_auth_util_unittest.cc
@@ -0,0 +1,87 @@
+// Copyright (c) 2012 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.
+
+#include "chrome/common/net/gaia/gaia_auth_util.h"
+
+#include "testing/gtest/include/gtest/gtest.h"
+
+namespace gaia {
+
+TEST(GaiaAuthUtilTest, EmailAddressNoOp) {
+  const char lower_case[] = "user@what.com";
+  EXPECT_EQ(lower_case, CanonicalizeEmail(lower_case));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressIgnoreCaps) {
+  EXPECT_EQ(CanonicalizeEmail("user@what.com"),
+            CanonicalizeEmail("UsEr@what.com"));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressIgnoreDomainCaps) {
+  EXPECT_EQ(CanonicalizeEmail("user@what.com"),
+            CanonicalizeEmail("UsEr@what.COM"));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressRejectOneUsernameDot) {
+  EXPECT_NE(CanonicalizeEmail("u.ser@what.com"),
+            CanonicalizeEmail("UsEr@what.com"));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressMatchWithOneUsernameDot) {
+  EXPECT_EQ(CanonicalizeEmail("u.ser@what.com"),
+            CanonicalizeEmail("U.sEr@what.com"));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressIgnoreOneUsernameDot) {
+  EXPECT_EQ(CanonicalizeEmail("us.er@gmail.com"),
+            CanonicalizeEmail("UsEr@gmail.com"));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressIgnoreManyUsernameDots) {
+  EXPECT_EQ(CanonicalizeEmail("u.ser@gmail.com"),
+            CanonicalizeEmail("Us.E.r@gmail.com"));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressIgnoreConsecutiveUsernameDots) {
+  EXPECT_EQ(CanonicalizeEmail("use.r@gmail.com"),
+            CanonicalizeEmail("Us....E.r@gmail.com"));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressDifferentOnesRejected) {
+  EXPECT_NE(CanonicalizeEmail("who@what.com"),
+            CanonicalizeEmail("Us....E.r@what.com"));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressIgnorePlusSuffix) {
+  const char with_plus[] = "user+cc@what.com";
+  EXPECT_EQ(with_plus, CanonicalizeEmail(with_plus));
+}
+
+TEST(GaiaAuthUtilTest, EmailAddressIgnoreMultiPlusSuffix) {
+  const char multi_plus[] = "user+cc+bcc@what.com";
+  EXPECT_EQ(multi_plus, CanonicalizeEmail(multi_plus));
+}
+
+TEST(GaiaAuthUtilTest, CanonicalizeDomain) {
+  const char domain[] = "example.com";
+  EXPECT_EQ(domain, CanonicalizeDomain("example.com"));
+  EXPECT_EQ(domain, CanonicalizeDomain("EXAMPLE.cOm"));
+}
+
+TEST(GaiaAuthUtilTest, ExtractDomainName) {
+  const char domain[] = "example.com";
+  EXPECT_EQ(domain, ExtractDomainName("who@example.com"));
+  EXPECT_EQ(domain, ExtractDomainName("who@EXAMPLE.cOm"));
+}
+
+TEST(GaiaAuthUtilTest, SanitizeMissingDomain) {
+  EXPECT_EQ("nodomain@gmail.com", SanitizeEmail("nodomain"));
+}
+
+TEST(GaiaAuthUtilTest, SanitizeExistingDomain) {
+  const char existing[] = "test@example.com";
+  EXPECT_EQ(existing, SanitizeEmail(existing));
+}
+
+}  // namespace gaia