| 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 | // This file contains one implementation of FormattedValue. |
| 9 | // Other independent implementations should go into their own cpp file for |
| 10 | // better dependency modularization. |
| 11 | |
| 12 | #include "unicode/ustring.h" |
| 13 | #include "formattedval_impl.h" |
| 14 | #include "number_types.h" |
| 15 | #include "formatted_string_builder.h" |
| 16 | #include "number_utils.h" |
| 17 | #include "static_unicode_sets.h" |
| 18 | |
| 19 | U_NAMESPACE_BEGIN |
| 20 | |
| 21 | |
| 22 | typedef FormattedStringBuilder::Field Field; |
| 23 | |
| 24 | |
| 25 | FormattedValueStringBuilderImpl::FormattedValueStringBuilderImpl(Field numericField) |
| 26 | : fNumericField(numericField) { |
| 27 | } |
| 28 | |
| 29 | FormattedValueStringBuilderImpl::~FormattedValueStringBuilderImpl() { |
| 30 | } |
| 31 | |
| 32 | |
| 33 | UnicodeString FormattedValueStringBuilderImpl::toString(UErrorCode&) const { |
| 34 | return fString.toUnicodeString(); |
| 35 | } |
| 36 | |
| 37 | UnicodeString FormattedValueStringBuilderImpl::toTempString(UErrorCode&) const { |
| 38 | return fString.toTempUnicodeString(); |
| 39 | } |
| 40 | |
| 41 | Appendable& FormattedValueStringBuilderImpl::appendTo(Appendable& appendable, UErrorCode&) const { |
| 42 | appendable.appendString(fString.chars(), fString.length()); |
| 43 | return appendable; |
| 44 | } |
| 45 | |
| 46 | UBool FormattedValueStringBuilderImpl::nextPosition(ConstrainedFieldPosition& cfpos, UErrorCode& status) const { |
| 47 | // NOTE: MSVC sometimes complains when implicitly converting between bool and UBool |
| 48 | return nextPositionImpl(cfpos, fNumericField, status) ? TRUE : FALSE; |
| 49 | } |
| 50 | |
| 51 | UBool FormattedValueStringBuilderImpl::nextFieldPosition(FieldPosition& fp, UErrorCode& status) const { |
| 52 | int32_t rawField = fp.getField(); |
| 53 | |
| 54 | if (rawField == FieldPosition::DONT_CARE) { |
| 55 | return FALSE; |
| 56 | } |
| 57 | |
| 58 | if (rawField < 0 || rawField >= UNUM_FIELD_COUNT) { |
| 59 | status = U_ILLEGAL_ARGUMENT_ERROR; |
| 60 | return FALSE; |
| 61 | } |
| 62 | |
| 63 | ConstrainedFieldPosition cfpos; |
| 64 | cfpos.constrainField(UFIELD_CATEGORY_NUMBER, rawField); |
| 65 | cfpos.setState(UFIELD_CATEGORY_NUMBER, rawField, fp.getBeginIndex(), fp.getEndIndex()); |
| 66 | if (nextPositionImpl(cfpos, 0, status)) { |
| 67 | fp.setBeginIndex(cfpos.getStart()); |
| 68 | fp.setEndIndex(cfpos.getLimit()); |
| 69 | return TRUE; |
| 70 | } |
| 71 | |
| 72 | // Special case: fraction should start after integer if fraction is not present |
| 73 | if (rawField == UNUM_FRACTION_FIELD && fp.getEndIndex() == 0) { |
| 74 | bool inside = false; |
| 75 | int32_t i = fString.fZero; |
| 76 | for (; i < fString.fZero + fString.fLength; i++) { |
| 77 | if (isIntOrGroup(fString.getFieldPtr()[i]) || fString.getFieldPtr()[i] == UNUM_DECIMAL_SEPARATOR_FIELD) { |
| 78 | inside = true; |
| 79 | } else if (inside) { |
| 80 | break; |
| 81 | } |
| 82 | } |
| 83 | fp.setBeginIndex(i - fString.fZero); |
| 84 | fp.setEndIndex(i - fString.fZero); |
| 85 | } |
| 86 | |
| 87 | return FALSE; |
| 88 | } |
| 89 | |
| 90 | void FormattedValueStringBuilderImpl::getAllFieldPositions(FieldPositionIteratorHandler& fpih, |
| 91 | UErrorCode& status) const { |
| 92 | ConstrainedFieldPosition cfpos; |
| 93 | while (nextPositionImpl(cfpos, 0, status)) { |
| 94 | fpih.addAttribute(cfpos.getField(), cfpos.getStart(), cfpos.getLimit()); |
| 95 | } |
| 96 | } |
| 97 | |
| 98 | // Signal the end of the string using a field that doesn't exist and that is |
| 99 | // different from UNUM_FIELD_COUNT, which is used for "null number field". |
| 100 | static constexpr Field kEndField = 0xff; |
| 101 | |
| 102 | bool FormattedValueStringBuilderImpl::nextPositionImpl(ConstrainedFieldPosition& cfpos, Field numericField, UErrorCode& /*status*/) const { |
| 103 | auto numericCAF = StringBuilderFieldUtils::expand(numericField); |
| 104 | int32_t fieldStart = -1; |
| 105 | Field currField = UNUM_FIELD_COUNT; |
| 106 | for (int32_t i = fString.fZero + cfpos.getLimit(); i <= fString.fZero + fString.fLength; i++) { |
| 107 | Field _field = (i < fString.fZero + fString.fLength) ? fString.getFieldPtr()[i] : kEndField; |
| 108 | // Case 1: currently scanning a field. |
| 109 | if (currField != UNUM_FIELD_COUNT) { |
| 110 | if (currField != _field) { |
| 111 | int32_t end = i - fString.fZero; |
| 112 | // Grouping separators can be whitespace; don't throw them out! |
| 113 | if (currField != UNUM_GROUPING_SEPARATOR_FIELD) { |
| 114 | end = trimBack(i - fString.fZero); |
| 115 | } |
| 116 | if (end <= fieldStart) { |
| 117 | // Entire field position is ignorable; skip. |
| 118 | fieldStart = -1; |
| 119 | currField = UNUM_FIELD_COUNT; |
| 120 | i--; // look at this index again |
| 121 | continue; |
| 122 | } |
| 123 | int32_t start = fieldStart; |
| 124 | if (currField != UNUM_GROUPING_SEPARATOR_FIELD) { |
| 125 | start = trimFront(start); |
| 126 | } |
| 127 | auto caf = StringBuilderFieldUtils::expand(currField); |
| 128 | cfpos.setState(caf.category, caf.field, start, end); |
| 129 | return true; |
| 130 | } |
| 131 | continue; |
| 132 | } |
| 133 | // Special case: coalesce the INTEGER if we are pointing at the end of the INTEGER. |
| 134 | if (cfpos.matchesField(UFIELD_CATEGORY_NUMBER, UNUM_INTEGER_FIELD) |
| 135 | && i > fString.fZero |
| 136 | // don't return the same field twice in a row: |
| 137 | && i - fString.fZero > cfpos.getLimit() |
| 138 | && isIntOrGroup(fString.getFieldPtr()[i - 1]) |
| 139 | && !isIntOrGroup(_field)) { |
| 140 | int j = i - 1; |
| 141 | for (; j >= fString.fZero && isIntOrGroup(fString.getFieldPtr()[j]); j--) {} |
| 142 | cfpos.setState( |
| 143 | UFIELD_CATEGORY_NUMBER, |
| 144 | UNUM_INTEGER_FIELD, |
| 145 | j - fString.fZero + 1, |
| 146 | i - fString.fZero); |
| 147 | return true; |
| 148 | } |
| 149 | // Special case: coalesce NUMERIC if we are pointing at the end of the NUMERIC. |
| 150 | if (numericField != 0 |
| 151 | && cfpos.matchesField(numericCAF.category, numericCAF.field) |
| 152 | && i > fString.fZero |
| 153 | // don't return the same field twice in a row: |
| 154 | && (i - fString.fZero > cfpos.getLimit() |
| 155 | || cfpos.getCategory() != numericCAF.category |
| 156 | || cfpos.getField() != numericCAF.field) |
| 157 | && isNumericField(fString.getFieldPtr()[i - 1]) |
| 158 | && !isNumericField(_field)) { |
| 159 | int j = i - 1; |
| 160 | for (; j >= fString.fZero && isNumericField(fString.getFieldPtr()[j]); j--) {} |
| 161 | cfpos.setState( |
| 162 | numericCAF.category, |
| 163 | numericCAF.field, |
| 164 | j - fString.fZero + 1, |
| 165 | i - fString.fZero); |
| 166 | return true; |
| 167 | } |
| 168 | // Special case: skip over INTEGER; will be coalesced later. |
| 169 | if (_field == UNUM_INTEGER_FIELD) { |
| 170 | _field = UNUM_FIELD_COUNT; |
| 171 | } |
| 172 | // Case 2: no field starting at this position. |
| 173 | if (_field == UNUM_FIELD_COUNT || _field == kEndField) { |
| 174 | continue; |
| 175 | } |
| 176 | // Case 3: check for field starting at this position |
| 177 | auto caf = StringBuilderFieldUtils::expand(_field); |
| 178 | if (cfpos.matchesField(caf.category, caf.field)) { |
| 179 | fieldStart = i - fString.fZero; |
| 180 | currField = _field; |
| 181 | } |
| 182 | } |
| 183 | |
| 184 | U_ASSERT(currField == UNUM_FIELD_COUNT); |
| 185 | return false; |
| 186 | } |
| 187 | |
| 188 | bool FormattedValueStringBuilderImpl::isIntOrGroup(Field field) { |
| 189 | return field == UNUM_INTEGER_FIELD |
| 190 | || field == UNUM_GROUPING_SEPARATOR_FIELD; |
| 191 | } |
| 192 | |
| 193 | bool FormattedValueStringBuilderImpl::isNumericField(Field field) { |
| 194 | return StringBuilderFieldUtils::isNumericField(field); |
| 195 | } |
| 196 | |
| 197 | int32_t FormattedValueStringBuilderImpl::trimBack(int32_t limit) const { |
| 198 | return unisets::get(unisets::DEFAULT_IGNORABLES)->spanBack( |
| 199 | fString.getCharPtr() + fString.fZero, |
| 200 | limit, |
| 201 | USET_SPAN_CONTAINED); |
| 202 | } |
| 203 | |
| 204 | int32_t FormattedValueStringBuilderImpl::trimFront(int32_t start) const { |
| 205 | return start + unisets::get(unisets::DEFAULT_IGNORABLES)->span( |
| 206 | fString.getCharPtr() + fString.fZero + start, |
| 207 | fString.fLength - start, |
| 208 | USET_SPAN_CONTAINED); |
| 209 | } |
| 210 | |
| 211 | |
| 212 | U_NAMESPACE_END |
| 213 | |
| 214 | #endif /* #if !UCONFIG_NO_FORMATTING */ |
| 215 | |