autofill: switch autofill_regexes to RE2 library

components/autofill/core/browser/autofill_regex_constants.cc

 // Copyright 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.
 
 // This file contains UTF8 strings that we want as char arrays.  To avoid
 // different compilers, we use a script to convert the UTF8 strings into
 // numeric literals (\x##).
 
 #include "components/autofill/core/browser/autofill_regex_constants.h"
 
+// This macro is to workaround the fact that RE2 library only supports ASCII
+// word boundaries and it is supposed to be the same as \b.
+#define WORDBREAK "(\\A|\\z|\\PL)"
+
 namespace autofill {
 
 /////////////////////////////////////////////////////////////////////////////
 // address_field.cc
 /////////////////////////////////////////////////////////////////////////////
 const char kAttentionIgnoredRe[] = "attention|attn";
 const char kRegionIgnoredRe[] =
     "province|region|other"
     "|provincia"  // es
     "|bairro|suburb";  // pt-BR, pt-PT
@@ skipping 56 matching lines @@
     "|país|pais"  // es
     "|国"  // ja-JP
     "|国家"  // zh-CN
     "|국가|나라";  // ko-KR
 const char kCountryLocationRe[] =
     "location";
 const char kZipCodeRe[] =
     "zip|postal|post.*code|pcode"
     "|pin.?code"  // en-IN
     "|postleitzahl"  // de-DE
-    "|\\bcp\\b"  // es
-    "|\\bcdp\\b"  // fr-FR
-    "|\\bcap\\b"  // it-IT
+    "|" WORDBREAK "cp" WORDBREAK  // es
+    "|" WORDBREAK "cdp" WORDBREAK  // fr-FR
+    "|" WORDBREAK "cap" WORDBREAK  // it-IT
     "|郵便番号"  // ja-JP
-    "|codigo|codpos|\\bcep\\b"  // pt-BR, pt-PT
+    "|codigo|codpos|" WORDBREAK "cep" WORDBREAK  // pt-BR, pt-PT
     "|Почтовый.?Индекс"  // ru
     "|邮政编码|邮编"  // zh-CN
     "|郵遞區號"  // zh-TW
     "|우편.?번호";  // ko-KR
 const char kZip4Re[] =
     "zip|^-$|post2"
     "|codpos2";  // pt-BR, pt-PT
 const char kCityRe[] =
     "city|town"
-    "|\\bort\\b|stadt"  // de-DE
+    "|" WORDBREAK "ort" WORDBREAK "|stadt"  // de-DE
     "|suburb"  // en-AU
     "|ciudad|provincia|localidad|poblacion"  // es
     "|ville|commune"  // fr-FR
     "|localita"  // it-IT
     "|市区町村"  // ja-JP
     "|cidade"  // pt-BR, pt-PT
     "|Город"  // ru
     "|市"  // zh-CN
     "|分區"  // zh-TW
     "|^시[^도·・]|시[·・]?군[·・]?구";  // ko-KR
 const char kStateRe[] =
-    "(?<!united )state|county|region|province"
+    "state|county|region|province"
     "|land"  // de-DE
     "|county|principality"  // en-UK
     "|都道府県"  // ja-JP
     "|estado|provincia"  // pt-BR, pt-PT
     "|область"  // ru
     "|省"  // zh-CN
     "|地區"  // zh-TW
     "|^시[·・]?도";  // ko-KR
 
 /////////////////////////////////////////////////////////////////////////////
 // credit_card_field.cc
 /////////////////////////////////////////////////////////////////////////////
 const char kNameOnCardRe[] =
-    "card.?(holder|owner)|name.*\\bon\\b.*card|(card|cc).?name|cc.?full.?name"
+    "card.?(holder|owner)|name.*" WORDBREAK "on" WORDBREAK ".*card"
+    "|(card|cc).?name|cc.?full.?name"
     "|karteninhaber"  // de-DE
     "|nombre.*tarjeta"  // es
     "|nom.*carte"  // fr-FR
     "|nome.*cart"  // it-IT
     "|名前"  // ja-JP
     "|Имя.*карты"  // ru
     "|信用卡开户名|开户名|持卡人姓名"  // zh-CN
     "|持卡人姓名";  // zh-TW
 const char kNameOnCardContextualRe[] =
     "name";
 const char kCardNumberRe[] =
     "(card|cc|acct).?(number|#|no|num)"
     "|nummer"  // de-DE
     "|credito|numero|número"  // es
     "|numéro"  // fr-FR
     "|カード番号"  // ja-JP
     "|Номер.*карты"  // ru
     "|信用卡号|信用卡号码"  // zh-CN
     "|信用卡卡號"  // zh-TW
     "|카드";  // ko-KR
 const char kCardCvcRe[] =
     "verification|card identification|security code|card code"
     "|cvn|cvv|cvc|csc|cvd|cid|ccv"
-    "|\\bcid\\b";
+    "|" WORDBREAK "cid" WORDBREAK;
 
 // "Expiration date" is the most common label here, but some pages have
 // "Expires", "exp. date" or "exp. month" and "exp. year".  We also look
 // for the field names ccmonth and ccyear, which appear on at least 4 of
 // our test pages.
 
 // On at least one page (The China Shop2.html) we find only the labels
 // "month" and "year".  So for now we match these words directly; we'll
 // see if this turns out to be too general.
 
@@ skipping 78 matching lines @@
     "|^nome";  // pt-BR, pt-PT
 const char kFirstNameRe[] =
     "first.*name|initials|fname|first$|given.*name"
     "|vorname"  // de-DE
     "|nombre"  // es
     "|forename|prénom|prenom"  // fr-FR
     "|名"  // ja-JP
     "|nome"  // pt-BR, pt-PT
     "|Имя"  // ru
     "|이름";  // ko-KR
-const char kMiddleInitialRe[] = "middle.*initial|m\\.i\\.|mi$|\\bmi\\b";
+const char kMiddleInitialRe[] =
+    "middle.*initial|m\\.i\\.|mi$|" WORDBREAK "mi" WORDBREAK;
 const char kMiddleNameRe[] =
     "middle.*name|mname|middle$"
     "|apellido.?materno|lastlastname";  // es
 const char kLastNameRe[] =
     "last.*name|lname|surname|last$|secondname|family.*name"
     "|nachname"  // de-DE
     "|apellido"  // es
     "|famille|^nom"  // fr-FR
     "|cognome"  // it-IT
     "|姓"  // ja-JP
@@ skipping 25 matching lines @@
     "^-$|^\\)$";
 const char kPhoneSuffixSeparatorRe[] =
     "^-$";
 const char kPhonePrefixRe[] =
     "prefix|exchange"
     "|preselection"  // fr-FR
     "|ddd";  // pt-BR, pt-PT
 const char kPhoneSuffixRe[] =
     "suffix";
 const char kPhoneExtensionRe[] =
-    "\\bext|ext\\b|extension"
+    WORDBREAK "ext|ext" WORDBREAK "|extension"
     "|ramal";  // pt-BR, pt-PT
 
 }  // namespace autofill
+
+#undef WORDBREAK