Remove diacritics when normalizing autofill profile strings for comparison.

components/autofill/core/browser/autofill_profile.cc

Index: components/autofill/core/browser/autofill_profile.cc
diff --git a/components/autofill/core/browser/autofill_profile.cc b/components/autofill/core/browser/autofill_profile.cc
index 0f69602a9265b7a33035cbcbb7552847a1b150ad..fe62770b51c5f28e0b354ae71a60aba30b0d2a47 100644
--- a/components/autofill/core/browser/autofill_profile.cc
+++ b/components/autofill/core/browser/autofill_profile.cc
@@ -36,6 +36,8 @@
 #include "components/autofill/core/common/form_field_data.h"
 #include "grit/components_strings.h"
 #include "third_party/icu/source/common/unicode/uchar.h"
+#include "third_party/icu/source/common/unicode/utypes.h"
+#include "third_party/icu/source/i18n/unicode/translit.h"
 #include "third_party/libaddressinput/chromium/addressinput_util.h"
 #include "third_party/libaddressinput/src/cpp/include/libaddressinput/address_data.h"
 #include "third_party/libaddressinput/src/cpp/include/libaddressinput/address_formatter.h"
@@ -235,6 +237,26 @@ class FindByPhone {
   std::string app_locale_;
 };
 
+base::string16 NormalizeForComparison(const base::string16& text) {
+  using icu::UnicodeString;
+  using icu::Transliterator;
+
+   // Remove diacritics and fold case.

[Mathieu, 2016/05/27 17:47:13]

extra space before comment

[Roger McFarlane (Chromium), 2016/05/27 20:20:42]

Done.

+  UErrorCode status = U_ZERO_ERROR;
+  std::unique_ptr<Transliterator> transliterator(Transliterator::createInstance(
+      "NFD; [:Nonspacing Mark:] Remove; Lower; NFC", UTRANS_FORWARD, status));

[Mathieu, 2016/05/27 17:47:13]

Can we put an explanation (or a link to the explanation) for the magic string?
The more details the better, so that the reader (which could be us in 6 months)
know what's happening.

[Roger McFarlane (Chromium), 2016/05/27 20:20:43]

Done.

+  if (U_FAILURE(status)) {
+    // This should not happen. Log the error and fall back.
+    LOG(ERROR) << "normalization failed: " << u_errorName(status);

[Mathieu, 2016/05/27 17:47:13]

Let's capitalize the comment and make it more explicit so that it's clear to
someone reading the logs.

[Roger McFarlane (Chromium), 2016/05/27 20:20:43]

Done.

+    return text;
+  }
+
+  UnicodeString value = UnicodeString(text.data(), text.length());
+  transliterator->transliterate(value);

[Mathieu, 2016/05/27 17:47:13]

can we check "if (transliterator)" here? ICU has a bad habit of returning null
instances sometimes.

[Roger McFarlane (Chromium), 2016/05/27 20:20:43]

Done.

+
+  return base::string16(value.getBuffer(), value.length());
+}
+
 }  // namespace
 
 AutofillProfile::AutofillProfile(const std::string& guid,
@@ -846,7 +868,7 @@ base::string16 AutofillProfile::CanonicalizeProfileString(
   if (previous_was_whitespace)
     ret.resize(ret.size() - 1);
 
-  return ret;
+  return NormalizeForComparison(ret);
 }
 
 // static