base/i18n/string_search.cc

   1 // Copyright (c) 2011 The Chromium Authors. All rights reserved.
   2 // Use of this source code is governed by a BSD-style license that can be
   3 // found in the LICENSE file.
   4
   5 #include "base/i18n/string_search.h"
   6 #include "base/logging.h"
   7
   8 #include "third_party/icu/source/i18n/unicode/usearch.h"
   9
  10 namespace base {
  11 namespace i18n {
  12
  13 FixedPatternStringSearchIgnoringCaseAndAccents::
  14 FixedPatternStringSearchIgnoringCaseAndAccents(const string16& find_this)
  15     : find_this_(find_this) {
  16   // usearch_open requires a valid string argument to be searched, even if we
  17   // want to set it by usearch_setText afterwards. So, supplying a dummy text.
  18   const string16& dummy = find_this_;
  19
  20   UErrorCode status = U_ZERO_ERROR;
  21   search_ = usearch_open(find_this_.data(), find_this_.size(),
  22                          dummy.data(), dummy.size(),
  23                          uloc_getDefault(),
  24                          NULL,  // breakiter
  25                          &status);
  26   if (U_SUCCESS(status)) {
  27     UCollator* collator = usearch_getCollator(search_);
  28     ucol_setStrength(collator, UCOL_PRIMARY);
  29     usearch_reset(search_);
  30   }
  31 }
  32
  33 FixedPatternStringSearchIgnoringCaseAndAccents::
  34 ~FixedPatternStringSearchIgnoringCaseAndAccents() {
  35   if (search_)
  36     usearch_close(search_);
  37 }
  38
  39 bool FixedPatternStringSearchIgnoringCaseAndAccents::Search(
  40     const string16& in_this, size_t* match_index, size_t* match_length) {
  41   UErrorCode status = U_ZERO_ERROR;
  42   usearch_setText(search_, in_this.data(), in_this.size(), &status);
  43
  44   // Default to basic substring search if usearch fails. According to
  45   // http://icu-project.org/apiref/icu4c/usearch_8h.html, usearch_open will fail
  46   // if either |find_this| or |in_this| are empty. In either case basic
  47   // substring search will give the correct return value.
  48   if (!U_SUCCESS(status)) {
  49     size_t index = in_this.find(find_this_);
  50     if (index == string16::npos) {
  51       return false;
  52     } else {
  53       if (match_index)
  54         *match_index = index;
  55       if (match_length)
  56         *match_length = find_this_.size();
  57       return true;
  58     }
  59   }
  60
  61   int32_t index = usearch_first(search_, &status);
  62   if (!U_SUCCESS(status) || index == USEARCH_DONE)
  63     return false;
  64   if (match_index)
  65     *match_index = static_cast<size_t>(index);
  66   if (match_length)
  67     *match_length = static_cast<size_t>(usearch_getMatchedLength(search_));
  68   return true;
  69 }
  70
  71 bool StringSearchIgnoringCaseAndAccents(const string16& find_this,
  72                                         const string16& in_this,
  73                                         size_t* match_index,
  74                                         size_t* match_length) {
  75   return FixedPatternStringSearchIgnoringCaseAndAccents(find_this).Search(
  76       in_this, match_index, match_length);
  77 }
  78
  79 }  // namespace i18n
  80 }  // namespace base