/src/mozilla-central/intl/icu/source/i18n/numparse_scientific.cpp
Line | Count | Source (jump to first uncovered line) |
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_scientific.h" |
14 | | #include "static_unicode_sets.h" |
15 | | |
16 | | using namespace icu; |
17 | | using namespace icu::numparse; |
18 | | using namespace icu::numparse::impl; |
19 | | |
20 | | |
21 | | namespace { |
22 | | |
23 | 0 | inline const UnicodeSet& minusSignSet() { |
24 | 0 | return *unisets::get(unisets::MINUS_SIGN); |
25 | 0 | } |
26 | | |
27 | 0 | inline const UnicodeSet& plusSignSet() { |
28 | 0 | return *unisets::get(unisets::PLUS_SIGN); |
29 | 0 | } |
30 | | |
31 | | } // namespace |
32 | | |
33 | | |
34 | | ScientificMatcher::ScientificMatcher(const DecimalFormatSymbols& dfs, const Grouper& grouper) |
35 | | : fExponentSeparatorString(dfs.getConstSymbol(DecimalFormatSymbols::kExponentialSymbol)), |
36 | 0 | fExponentMatcher(dfs, grouper, PARSE_FLAG_INTEGER_ONLY | PARSE_FLAG_GROUPING_DISABLED) { |
37 | 0 |
|
38 | 0 | const UnicodeString& minusSign = dfs.getConstSymbol(DecimalFormatSymbols::kMinusSignSymbol); |
39 | 0 | if (minusSignSet().contains(minusSign)) { |
40 | 0 | fCustomMinusSign.setToBogus(); |
41 | 0 | } else { |
42 | 0 | fCustomMinusSign = minusSign; |
43 | 0 | } |
44 | 0 |
|
45 | 0 | const UnicodeString& plusSign = dfs.getConstSymbol(DecimalFormatSymbols::kPlusSignSymbol); |
46 | 0 | if (plusSignSet().contains(plusSign)) { |
47 | 0 | fCustomPlusSign.setToBogus(); |
48 | 0 | } else { |
49 | 0 | fCustomPlusSign = plusSign; |
50 | 0 | } |
51 | 0 | } |
52 | | |
53 | 0 | bool ScientificMatcher::match(StringSegment& segment, ParsedNumber& result, UErrorCode& status) const { |
54 | 0 | // Only accept scientific notation after the mantissa. |
55 | 0 | if (!result.seenNumber()) { |
56 | 0 | return false; |
57 | 0 | } |
58 | 0 | |
59 | 0 | // First match the scientific separator, and then match another number after it. |
60 | 0 | // NOTE: This is guarded by the smoke test; no need to check fExponentSeparatorString length again. |
61 | 0 | int overlap1 = segment.getCommonPrefixLength(fExponentSeparatorString); |
62 | 0 | if (overlap1 == fExponentSeparatorString.length()) { |
63 | 0 | // Full exponent separator match. |
64 | 0 |
|
65 | 0 | // First attempt to get a code point, returning true if we can't get one. |
66 | 0 | if (segment.length() == overlap1) { |
67 | 0 | return true; |
68 | 0 | } |
69 | 0 | segment.adjustOffset(overlap1); |
70 | 0 |
|
71 | 0 | // Allow a sign, and then try to match digits. |
72 | 0 | int8_t exponentSign = 1; |
73 | 0 | if (segment.startsWith(minusSignSet())) { |
74 | 0 | exponentSign = -1; |
75 | 0 | segment.adjustOffsetByCodePoint(); |
76 | 0 | } else if (segment.startsWith(plusSignSet())) { |
77 | 0 | segment.adjustOffsetByCodePoint(); |
78 | 0 | } else if (segment.startsWith(fCustomMinusSign)) { |
79 | 0 | // Note: call site is guarded with startsWith, which returns false on empty string |
80 | 0 | int32_t overlap2 = segment.getCommonPrefixLength(fCustomMinusSign); |
81 | 0 | if (overlap2 != fCustomMinusSign.length()) { |
82 | 0 | // Partial custom sign match; un-match the exponent separator. |
83 | 0 | segment.adjustOffset(-overlap1); |
84 | 0 | return true; |
85 | 0 | } |
86 | 0 | exponentSign = -1; |
87 | 0 | segment.adjustOffset(overlap2); |
88 | 0 | } else if (segment.startsWith(fCustomPlusSign)) { |
89 | 0 | // Note: call site is guarded with startsWith, which returns false on empty string |
90 | 0 | int32_t overlap2 = segment.getCommonPrefixLength(fCustomPlusSign); |
91 | 0 | if (overlap2 != fCustomPlusSign.length()) { |
92 | 0 | // Partial custom sign match; un-match the exponent separator. |
93 | 0 | segment.adjustOffset(-overlap1); |
94 | 0 | return true; |
95 | 0 | } |
96 | 0 | segment.adjustOffset(overlap2); |
97 | 0 | } |
98 | 0 |
|
99 | 0 | // We are supposed to accept E0 after NaN, so we need to make sure result.quantity is available. |
100 | 0 | bool wasBogus = result.quantity.bogus; |
101 | 0 | result.quantity.bogus = false; |
102 | 0 | int digitsOffset = segment.getOffset(); |
103 | 0 | bool digitsReturnValue = fExponentMatcher.match(segment, result, exponentSign, status); |
104 | 0 | result.quantity.bogus = wasBogus; |
105 | 0 |
|
106 | 0 | if (segment.getOffset() != digitsOffset) { |
107 | 0 | // At least one exponent digit was matched. |
108 | 0 | result.flags |= FLAG_HAS_EXPONENT; |
109 | 0 | } else { |
110 | 0 | // No exponent digits were matched; un-match the exponent separator. |
111 | 0 | segment.adjustOffset(-overlap1); |
112 | 0 | } |
113 | 0 | return digitsReturnValue; |
114 | 0 |
|
115 | 0 | } else if (overlap1 == segment.length()) { |
116 | 0 | // Partial exponent separator match |
117 | 0 | return true; |
118 | 0 | } |
119 | 0 | |
120 | 0 | // No match |
121 | 0 | return false; |
122 | 0 | } |
123 | | |
124 | 0 | bool ScientificMatcher::smokeTest(const StringSegment& segment) const { |
125 | 0 | return segment.startsWith(fExponentSeparatorString); |
126 | 0 | } |
127 | | |
128 | 0 | UnicodeString ScientificMatcher::toString() const { |
129 | 0 | return u"<Scientific>"; |
130 | 0 | } |
131 | | |
132 | | |
133 | | #endif /* #if !UCONFIG_NO_FORMATTING */ |