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 <stdint.h>
6
7 #include "base/i18n/string_search.h"
8 #include "base/logging.h"
9
10 #include "third_party/icu/source/i18n/unicode/usearch.h"
11
12 namespace base {
13 namespace i18n {
14
15 FixedPatternStringSearchIgnoringCaseAndAccents::
FixedPatternStringSearchIgnoringCaseAndAccents(const string16 & find_this)16 FixedPatternStringSearchIgnoringCaseAndAccents(const string16& find_this)
17 : find_this_(find_this) {
18 // usearch_open requires a valid string argument to be searched, even if we
19 // want to set it by usearch_setText afterwards. So, supplying a dummy text.
20 const string16& dummy = find_this_;
21
22 UErrorCode status = U_ZERO_ERROR;
23 search_ = usearch_open(find_this_.data(), find_this_.size(), dummy.data(),
24 dummy.size(), uloc_getDefault(),
25 nullptr, // breakiter
26 &status);
27 if (U_SUCCESS(status)) {
28 UCollator* collator = usearch_getCollator(search_);
29 ucol_setStrength(collator, UCOL_PRIMARY);
30 usearch_reset(search_);
31 }
32 }
33
34 FixedPatternStringSearchIgnoringCaseAndAccents::
~FixedPatternStringSearchIgnoringCaseAndAccents()35 ~FixedPatternStringSearchIgnoringCaseAndAccents() {
36 if (search_)
37 usearch_close(search_);
38 }
39
Search(const string16 & in_this,size_t * match_index,size_t * match_length)40 bool FixedPatternStringSearchIgnoringCaseAndAccents::Search(
41 const string16& in_this, size_t* match_index, size_t* match_length) {
42 UErrorCode status = U_ZERO_ERROR;
43 usearch_setText(search_, in_this.data(), in_this.size(), &status);
44
45 // Default to basic substring search if usearch fails. According to
46 // http://icu-project.org/apiref/icu4c/usearch_8h.html, usearch_open will fail
47 // if either |find_this| or |in_this| are empty. In either case basic
48 // substring search will give the correct return value.
49 if (!U_SUCCESS(status)) {
50 size_t index = in_this.find(find_this_);
51 if (index == string16::npos) {
52 return false;
53 } else {
54 if (match_index)
55 *match_index = index;
56 if (match_length)
57 *match_length = find_this_.size();
58 return true;
59 }
60 }
61
62 int32_t index = usearch_first(search_, &status);
63 if (!U_SUCCESS(status) || index == USEARCH_DONE)
64 return false;
65 if (match_index)
66 *match_index = static_cast<size_t>(index);
67 if (match_length)
68 *match_length = static_cast<size_t>(usearch_getMatchedLength(search_));
69 return true;
70 }
71
StringSearchIgnoringCaseAndAccents(const string16 & find_this,const string16 & in_this,size_t * match_index,size_t * match_length)72 bool StringSearchIgnoringCaseAndAccents(const string16& find_this,
73 const string16& in_this,
74 size_t* match_index,
75 size_t* match_length) {
76 return FixedPatternStringSearchIgnoringCaseAndAccents(find_this).Search(
77 in_this, match_index, match_length);
78 }
79
80 } // namespace i18n
81 } // namespace base
82