]>
Commit | Line | Data |
---|---|---|
0f5d89e8 A |
1 | // © 2018 and later: Unicode, Inc. and others. |
2 | // License & terms of use: http://www.unicode.org/copyright.html | |
3 | ||
4 | #include "unicode/utypes.h" | |
5 | ||
6 | #if !UCONFIG_NO_FORMATTING | |
7 | ||
8 | // Allow implicit conversion from char16_t* to UnicodeString for this file: | |
9 | // Helpful in toString methods and elsewhere. | |
10 | #define UNISTR_FROM_STRING_EXPLICIT | |
11 | ||
12 | #include "numparse_types.h" | |
13 | #include "numparse_currency.h" | |
14 | #include "ucurrimp.h" | |
15 | #include "unicode/errorcode.h" | |
16 | #include "numparse_utils.h" | |
17 | ||
18 | using namespace icu; | |
19 | using namespace icu::numparse; | |
20 | using namespace icu::numparse::impl; | |
21 | ||
22 | ||
23 | CombinedCurrencyMatcher::CombinedCurrencyMatcher(const CurrencySymbols& currencySymbols, const DecimalFormatSymbols& dfs, | |
24 | parse_flags_t parseFlags, UErrorCode& status) | |
25 | : fCurrency1(currencySymbols.getCurrencySymbol(status)), | |
26 | fCurrency2(currencySymbols.getIntlCurrencySymbol(status)), | |
27 | fUseFullCurrencyData(0 == (parseFlags & PARSE_FLAG_NO_FOREIGN_CURRENCY)), | |
c5116b9f | 28 | fCurrencyTrails(0 != (parseFlags & PARSE_FLAG_HAS_TRAIL_CURRENCY)), // Apple <rdar://problem/51938595> |
0f5d89e8 A |
29 | afterPrefixInsert(dfs.getPatternForCurrencySpacing(UNUM_CURRENCY_INSERT, false, status)), |
30 | beforeSuffixInsert(dfs.getPatternForCurrencySpacing(UNUM_CURRENCY_INSERT, true, status)), | |
31 | fLocaleName(dfs.getLocale().getName(), -1, status) { | |
32 | utils::copyCurrencyCode(fCurrencyCode, currencySymbols.getIsoCode()); | |
33 | ||
34 | // Pre-load the long names for the current locale and currency | |
35 | // if we are parsing without the full currency data. | |
36 | if (!fUseFullCurrencyData) { | |
37 | for (int32_t i=0; i<StandardPlural::COUNT; i++) { | |
38 | auto plural = static_cast<StandardPlural::Form>(i); | |
39 | fLocalLongNames[i] = currencySymbols.getPluralName(plural, status); | |
40 | } | |
41 | } | |
42 | ||
43 | // TODO: Figure out how to make this faster and re-enable. | |
44 | // Computing the "lead code points" set for fastpathing is too slow to use in production. | |
45 | // See http://bugs.icu-project.org/trac/ticket/13584 | |
46 | // // Compute the full set of characters that could be the first in a currency to allow for | |
47 | // // efficient smoke test. | |
48 | // fLeadCodePoints.add(fCurrency1.char32At(0)); | |
49 | // fLeadCodePoints.add(fCurrency2.char32At(0)); | |
50 | // fLeadCodePoints.add(beforeSuffixInsert.char32At(0)); | |
51 | // uprv_currencyLeads(fLocaleName.data(), fLeadCodePoints, status); | |
52 | // // Always apply case mapping closure for currencies | |
53 | // fLeadCodePoints.closeOver(USET_ADD_CASE_MAPPINGS); | |
54 | // fLeadCodePoints.freeze(); | |
55 | } | |
56 | ||
57 | bool | |
58 | CombinedCurrencyMatcher::match(StringSegment& segment, ParsedNumber& result, UErrorCode& status) const { | |
59 | if (result.currencyCode[0] != 0) { | |
60 | return false; | |
61 | } | |
62 | ||
63 | // Try to match a currency spacing separator. | |
64 | int32_t initialOffset = segment.getOffset(); | |
65 | bool maybeMore = false; | |
c5116b9f A |
66 | if (result.seenNumber() && !beforeSuffixInsert.isEmpty() |
67 | && segment.length() != 0) { // Apple <rdar://problem/51938595> | |
0f5d89e8 A |
68 | int32_t overlap = segment.getCommonPrefixLength(beforeSuffixInsert); |
69 | if (overlap == beforeSuffixInsert.length()) { | |
70 | segment.adjustOffset(overlap); | |
71 | // Note: let currency spacing be a weak match. Don't update chars consumed. | |
72 | } | |
73 | maybeMore = maybeMore || overlap == segment.length(); | |
74 | } | |
75 | ||
76 | // Match the currency string, and reset if we didn't find one. | |
77 | maybeMore = maybeMore || matchCurrency(segment, result, status); | |
78 | if (result.currencyCode[0] == 0) { | |
79 | segment.setOffset(initialOffset); | |
80 | return maybeMore; | |
81 | } | |
82 | ||
83 | // Try to match a currency spacing separator. | |
c5116b9f A |
84 | if (!result.seenNumber() && !afterPrefixInsert.isEmpty() |
85 | && segment.length() != 0) { // Apple <rdar://problem/51938595> | |
0f5d89e8 A |
86 | int32_t overlap = segment.getCommonPrefixLength(afterPrefixInsert); |
87 | if (overlap == afterPrefixInsert.length()) { | |
88 | segment.adjustOffset(overlap); | |
89 | // Note: let currency spacing be a weak match. Don't update chars consumed. | |
90 | } | |
91 | maybeMore = maybeMore || overlap == segment.length(); | |
92 | } | |
93 | ||
94 | return maybeMore; | |
95 | } | |
96 | ||
97 | bool CombinedCurrencyMatcher::matchCurrency(StringSegment& segment, ParsedNumber& result, | |
98 | UErrorCode& status) const { | |
99 | bool maybeMore = false; | |
100 | ||
101 | int32_t overlap1; | |
102 | if (!fCurrency1.isEmpty()) { | |
103 | overlap1 = segment.getCaseSensitivePrefixLength(fCurrency1); | |
c5116b9f | 104 | } else if (!fUseFullCurrencyData && (!fCurrencyTrails || result.seenNumber())) { // Apple <rdar://problem/46915356><rdar://problem/51938595> |
3d1f044b | 105 | overlap1 = 0; |
0f5d89e8 A |
106 | } else { |
107 | overlap1 = -1; | |
108 | } | |
109 | maybeMore = maybeMore || overlap1 == segment.length(); | |
110 | if (overlap1 == fCurrency1.length()) { | |
111 | utils::copyCurrencyCode(result.currencyCode, fCurrencyCode); | |
112 | segment.adjustOffset(overlap1); | |
113 | result.setCharsConsumed(segment); | |
114 | return maybeMore; | |
115 | } | |
116 | ||
117 | int32_t overlap2; | |
118 | if (!fCurrency2.isEmpty()) { | |
3d1f044b A |
119 | // ISO codes should be accepted case-insensitive. |
120 | // https://unicode-org.atlassian.net/browse/ICU-13696 | |
121 | overlap2 = segment.getCommonPrefixLength(fCurrency2); | |
0f5d89e8 A |
122 | } else { |
123 | overlap2 = -1; | |
124 | } | |
125 | maybeMore = maybeMore || overlap2 == segment.length(); | |
126 | if (overlap2 == fCurrency2.length()) { | |
127 | utils::copyCurrencyCode(result.currencyCode, fCurrencyCode); | |
128 | segment.adjustOffset(overlap2); | |
129 | result.setCharsConsumed(segment); | |
130 | return maybeMore; | |
131 | } | |
132 | ||
133 | if (fUseFullCurrencyData) { | |
134 | // Use the full currency data. | |
135 | // NOTE: This call site should be improved with #13584. | |
136 | const UnicodeString segmentString = segment.toTempUnicodeString(); | |
137 | ||
138 | // Try to parse the currency | |
139 | ParsePosition ppos(0); | |
140 | int32_t partialMatchLen = 0; | |
141 | uprv_parseCurrency( | |
142 | fLocaleName.data(), | |
143 | segmentString, | |
144 | ppos, | |
145 | UCURR_SYMBOL_NAME, // checks for both UCURR_SYMBOL_NAME and UCURR_LONG_NAME | |
146 | &partialMatchLen, | |
147 | result.currencyCode, | |
148 | status); | |
149 | maybeMore = maybeMore || partialMatchLen == segment.length(); | |
150 | ||
151 | if (U_SUCCESS(status) && ppos.getIndex() != 0) { | |
152 | // Complete match. | |
153 | // NOTE: The currency code should already be saved in the ParsedNumber. | |
154 | segment.adjustOffset(ppos.getIndex()); | |
155 | result.setCharsConsumed(segment); | |
156 | return maybeMore; | |
157 | } | |
158 | ||
159 | } else { | |
160 | // Use the locale long names. | |
161 | int32_t longestFullMatch = 0; | |
162 | for (int32_t i=0; i<StandardPlural::COUNT; i++) { | |
163 | const UnicodeString& name = fLocalLongNames[i]; | |
164 | int32_t overlap = segment.getCommonPrefixLength(name); | |
165 | if (overlap == name.length() && name.length() > longestFullMatch) { | |
166 | longestFullMatch = name.length(); | |
167 | } | |
168 | maybeMore = maybeMore || overlap > 0; | |
169 | } | |
170 | if (longestFullMatch > 0) { | |
171 | utils::copyCurrencyCode(result.currencyCode, fCurrencyCode); | |
172 | segment.adjustOffset(longestFullMatch); | |
173 | result.setCharsConsumed(segment); | |
174 | return maybeMore; | |
175 | } | |
176 | } | |
177 | ||
178 | // No match found. | |
179 | return maybeMore; | |
180 | } | |
181 | ||
182 | bool CombinedCurrencyMatcher::smokeTest(const StringSegment&) const { | |
183 | // TODO: See constructor | |
184 | return true; | |
185 | //return segment.startsWith(fLeadCodePoints); | |
186 | } | |
187 | ||
188 | UnicodeString CombinedCurrencyMatcher::toString() const { | |
189 | return u"<CombinedCurrencyMatcher>"; | |
190 | } | |
191 | ||
192 | ||
193 | #endif /* #if !UCONFIG_NO_FORMATTING */ |