1// Copyright (C) 2020 The Qt Company Ltd.
2// Copyright (C) 2013 Aleix Pol Gonzalez <aleixpol@kde.org>
3// SPDX-License-Identifier: LicenseRef-Qt-Commercial OR LGPL-3.0-only OR GPL-2.0-only OR GPL-3.0-only
4
5#include "qcollator_p.h"
6#include "qlocale_p.h"
7#include "qstringlist.h"
8#include "qstring.h"
9
10#include <unicode/utypes.h>
11#include <unicode/ucol.h>
12#include <unicode/ustring.h>
13#include <unicode/ures.h>
14
15#include "qdebug.h"
16
17QT_BEGIN_NAMESPACE
18
19void QCollatorPrivate::init()
20{
21 cleanup();
22 if (isC())
23 return;
24
25 UErrorCode status = U_ZERO_ERROR;
26 QByteArray name = QLocalePrivate::get(l: locale)->bcp47Name(separator: '_');
27 collator = ucol_open(loc: name.constData(), status: &status);
28 if (U_FAILURE(code: status)) {
29 qWarning(msg: "Could not create collator: %d", status);
30 collator = nullptr;
31 dirty = false;
32 return;
33 }
34
35 // enable normalization by default
36 ucol_setAttribute(coll: collator, attr: UCOL_NORMALIZATION_MODE, value: UCOL_ON, status: &status);
37
38 // The strength attribute in ICU is rather badly documented. Basically UCOL_PRIMARY
39 // ignores differences between base characters and accented characters as well as case.
40 // So A and A-umlaut would compare equal.
41 // UCOL_SECONDARY ignores case differences. UCOL_TERTIARY is the default in most languages
42 // and does case sensitive comparison.
43 // UCOL_QUATERNARY is used as default in a few languages such as Japanese to take care of some
44 // additional differences in those languages.
45 UColAttributeValue val = (caseSensitivity == Qt::CaseSensitive)
46 ? UCOL_DEFAULT_STRENGTH : UCOL_SECONDARY;
47
48 status = U_ZERO_ERROR;
49 ucol_setAttribute(coll: collator, attr: UCOL_STRENGTH, value: val, status: &status);
50 if (U_FAILURE(code: status))
51 qWarning(msg: "ucol_setAttribute: Case First failed: %d", status);
52
53 status = U_ZERO_ERROR;
54 ucol_setAttribute(coll: collator, attr: UCOL_NUMERIC_COLLATION, value: numericMode ? UCOL_ON : UCOL_OFF, status: &status);
55 if (U_FAILURE(code: status))
56 qWarning(msg: "ucol_setAttribute: numeric collation failed: %d", status);
57
58 status = U_ZERO_ERROR;
59 ucol_setAttribute(coll: collator, attr: UCOL_ALTERNATE_HANDLING,
60 value: ignorePunctuation ? UCOL_SHIFTED : UCOL_NON_IGNORABLE, status: &status);
61 if (U_FAILURE(code: status))
62 qWarning(msg: "ucol_setAttribute: Alternate handling failed: %d", status);
63
64 dirty = false;
65}
66
67void QCollatorPrivate::cleanup()
68{
69 if (collator)
70 ucol_close(coll: collator);
71 collator = nullptr;
72}
73
74int QCollator::compare(QStringView s1, QStringView s2) const
75{
76 if (!s1.size())
77 return s2.size() ? -1 : 0;
78 if (!s2.size())
79 return +1;
80
81 d->ensureInitialized();
82
83 if (d->collator) {
84 // truncating sizes (QTBUG-105038)
85 return ucol_strcoll(coll: d->collator,
86 source: reinterpret_cast<const UChar *>(s1.data()), sourceLength: s1.size(),
87 target: reinterpret_cast<const UChar *>(s2.data()), targetLength: s2.size());
88 }
89
90 return QtPrivate::compareStrings(lhs: s1, rhs: s2, cs: d->caseSensitivity);
91}
92
93QCollatorSortKey QCollator::sortKey(const QString &string) const
94{
95 d->ensureInitialized();
96
97 if (d->isC())
98 return QCollatorSortKey(new QCollatorSortKeyPrivate(string.toUtf8()));
99
100 if (d->collator) {
101 QByteArray result(16 + string.size() + (string.size() >> 2), Qt::Uninitialized);
102 // truncating sizes (QTBUG-105038)
103 int size = ucol_getSortKey(coll: d->collator, source: (const UChar *)string.constData(),
104 sourceLength: string.size(), result: (uint8_t *)result.data(), resultLength: result.size());
105 if (size > result.size()) {
106 result.resize(size);
107 size = ucol_getSortKey(coll: d->collator, source: (const UChar *)string.constData(),
108 sourceLength: string.size(), result: (uint8_t *)result.data(), resultLength: result.size());
109 }
110 result.truncate(pos: size);
111 return QCollatorSortKey(new QCollatorSortKeyPrivate(std::move(result)));
112 }
113
114 return QCollatorSortKey(new QCollatorSortKeyPrivate(QByteArray()));
115}
116
117int QCollatorSortKey::compare(const QCollatorSortKey &otherKey) const
118{
119 return qstrcmp(str1: d->m_key, str2: otherKey.d->m_key);
120}
121
122QT_END_NAMESPACE
123

source code of qtbase/src/corelib/text/qcollator_icu.cpp