OLD | NEW |
1 // Copyright 2011 the V8 project authors. All rights reserved. | 1 // Copyright 2011 the V8 project authors. All rights reserved. |
2 // Use of this source code is governed by a BSD-style license that can be | 2 // Use of this source code is governed by a BSD-style license that can be |
3 // found in the LICENSE file. | 3 // found in the LICENSE file. |
4 | 4 |
5 #ifndef V8_DATEPARSER_INL_H_ | 5 #ifndef V8_DATEPARSER_INL_H_ |
6 #define V8_DATEPARSER_INL_H_ | 6 #define V8_DATEPARSER_INL_H_ |
7 | 7 |
8 #include "src/dateparser.h" | 8 #include "src/dateparser.h" |
9 | 9 |
10 namespace v8 { | 10 namespace v8 { |
11 namespace internal { | 11 namespace internal { |
12 | 12 |
13 template <typename Char> | 13 template <typename Char> |
14 bool DateParser::Parse(Vector<Char> str, | 14 bool DateParser::Parse(Vector<Char> str, |
15 FixedArray* out, | 15 FixedArray* out, |
16 UnicodeCache* unicode_cache) { | 16 UnicodeCache* unicode_cache) { |
17 DCHECK(out->length() >= OUTPUT_SIZE); | 17 DCHECK(out->length() >= OUTPUT_SIZE); |
18 InputReader<Char> in(unicode_cache, str); | 18 InputReader<Char> in(unicode_cache, str); |
19 DateStringTokenizer<Char> scanner(&in); | 19 DateStringTokenizer<Char> scanner(&in); |
20 TimeZoneComposer tz; | 20 TimeZoneComposer tz; |
21 TimeComposer time; | 21 TimeComposer time; |
22 DayComposer day; | 22 DayComposer day; |
23 | 23 |
24 // Specification: | 24 // Specification: |
25 // Accept ES5 ISO 8601 date-time-strings or legacy dates compatible | 25 // Accept ES6 ISO 8601 date-time-strings or legacy dates compatible |
26 // with Safari. | 26 // with Safari. |
27 // ES5 ISO 8601 dates: | 27 // ES6 ISO 8601 dates: |
28 // [('-'|'+')yy]yyyy[-MM[-DD]][THH:mm[:ss[.sss]][Z|(+|-)hh:mm]] | 28 // [('-'|'+')yy]yyyy[-MM[-DD]][THH:mm[:ss[.sss]][Z|(+|-)hh:mm]] |
29 // where yyyy is in the range 0000..9999 and | 29 // where yyyy is in the range 0000..9999 and |
30 // +/-yyyyyy is in the range -999999..+999999 - | 30 // +/-yyyyyy is in the range -999999..+999999 - |
31 // but -000000 is invalid (year zero must be positive), | 31 // but -000000 is invalid (year zero must be positive), |
32 // MM is in the range 01..12, | 32 // MM is in the range 01..12, |
33 // DD is in the range 01..31, | 33 // DD is in the range 01..31, |
34 // MM and DD defaults to 01 if missing,, | 34 // MM and DD defaults to 01 if missing,, |
35 // HH is generally in the range 00..23, but can be 24 if mm, ss | 35 // HH is generally in the range 00..23, but can be 24 if mm, ss |
36 // and sss are zero (or missing), representing midnight at the | 36 // and sss are zero (or missing), representing midnight at the |
37 // end of a day, | 37 // end of a day, |
38 // mm and ss are in the range 00..59, | 38 // mm and ss are in the range 00..59, |
39 // sss is in the range 000..999, | 39 // sss is in the range 000..999, |
40 // hh is in the range 00..23, | 40 // hh is in the range 00..23, |
41 // mm, ss, and sss default to 00 if missing, and | 41 // mm, ss, and sss default to 00 if missing, and |
42 // timezone defaults to Z if missing | 42 // timezone defaults to local time if missing. |
43 // (following Safari, ISO actually demands local time). | |
44 // Extensions: | 43 // Extensions: |
45 // We also allow sss to have more or less than three digits (but at | 44 // We also allow sss to have more or less than three digits (but at |
46 // least one). | 45 // least one). |
47 // We allow hh:mm to be specified as hhmm. | 46 // We allow hh:mm to be specified as hhmm. |
48 // Legacy dates: | 47 // Legacy dates: |
49 // Any unrecognized word before the first number is ignored. | 48 // Any unrecognized word before the first number is ignored. |
50 // Parenthesized text is ignored. | 49 // Parenthesized text is ignored. |
51 // An unsigned number followed by ':' is a time value, and is | 50 // An unsigned number followed by ':' is a time value, and is |
52 // added to the TimeComposer. A number followed by '::' adds a second | 51 // added to the TimeComposer. A number followed by '::' adds a second |
53 // zero as well. A number followed by '.' is also a time and must be | 52 // zero as well. A number followed by '.' is also a time and must be |
54 // followed by milliseconds. | 53 // followed by milliseconds. |
55 // Any other number is a date component and is added to DayComposer. | 54 // Any other number is a date component and is added to DayComposer. |
56 // A month name (or really: any word having the same first three letters | 55 // A month name (or really: any word having the same first three letters |
57 // as a month name) is recorded as a named month in the Day composer. | 56 // as a month name) is recorded as a named month in the Day composer. |
58 // A word recognizable as a time-zone is recorded as such, as is | 57 // A word recognizable as a time-zone is recorded as such, as is |
59 // '(+|-)(hhmm|hh:)'. | 58 // '(+|-)(hhmm|hh:)'. |
60 // Legacy dates don't allow extra signs ('+' or '-') or umatched ')' | 59 // Legacy dates don't allow extra signs ('+' or '-') or umatched ')' |
61 // after a number has been read (before the first number, any garbage | 60 // after a number has been read (before the first number, any garbage |
62 // is allowed). | 61 // is allowed). |
63 // Intersection of the two: | 62 // Intersection of the two: |
64 // A string that matches both formats (e.g. 1970-01-01) will be | 63 // A string that matches both formats (e.g. 1970-01-01) will be |
65 // parsed as an ES5 date-time string - which means it will default | 64 // parsed as an ES6 date-time string. |
66 // to UTC time-zone. That's unavoidable if following the ES5 | 65 // After a valid "T" has been read while scanning an ES6 datetime string, |
67 // specification. | |
68 // After a valid "T" has been read while scanning an ES5 datetime string, | |
69 // the input can no longer be a valid legacy date, since the "T" is a | 66 // the input can no longer be a valid legacy date, since the "T" is a |
70 // garbage string after a number has been read. | 67 // garbage string after a number has been read. |
71 | 68 |
72 // First try getting as far as possible with as ES5 Date Time String. | 69 // First try getting as far as possible with as ES6 Date Time String. |
73 DateToken next_unhandled_token = ParseES5DateTime(&scanner, &day, &time, &tz); | 70 DateToken next_unhandled_token = ParseES6DateTime(&scanner, &day, &time, &tz); |
74 if (next_unhandled_token.IsInvalid()) return false; | 71 if (next_unhandled_token.IsInvalid()) return false; |
75 bool has_read_number = !day.IsEmpty(); | 72 bool has_read_number = !day.IsEmpty(); |
76 // If there's anything left, continue with the legacy parser. | 73 // If there's anything left, continue with the legacy parser. |
77 for (DateToken token = next_unhandled_token; | 74 for (DateToken token = next_unhandled_token; |
78 !token.IsEndOfInput(); | 75 !token.IsEndOfInput(); |
79 token = scanner.Next()) { | 76 token = scanner.Next()) { |
80 if (token.IsNumber()) { | 77 if (token.IsNumber()) { |
81 has_read_number = true; | 78 has_read_number = true; |
82 int n = token.number(); | 79 int n = token.number(); |
83 if (scanner.SkipSymbol(':')) { | 80 if (scanner.SkipSymbol(':')) { |
(...skipping 104 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
188 } | 185 } |
189 if (in_->SkipParentheses()) { | 186 if (in_->SkipParentheses()) { |
190 return DateToken::Unknown(); | 187 return DateToken::Unknown(); |
191 } | 188 } |
192 in_->Next(); | 189 in_->Next(); |
193 return DateToken::Unknown(); | 190 return DateToken::Unknown(); |
194 } | 191 } |
195 | 192 |
196 | 193 |
197 template <typename Char> | 194 template <typename Char> |
198 DateParser::DateToken DateParser::ParseES5DateTime( | 195 DateParser::DateToken DateParser::ParseES6DateTime( |
199 DateStringTokenizer<Char>* scanner, | 196 DateStringTokenizer<Char>* scanner, |
200 DayComposer* day, | 197 DayComposer* day, |
201 TimeComposer* time, | 198 TimeComposer* time, |
202 TimeZoneComposer* tz) { | 199 TimeZoneComposer* tz) { |
203 DCHECK(day->IsEmpty()); | 200 DCHECK(day->IsEmpty()); |
204 DCHECK(time->IsEmpty()); | 201 DCHECK(time->IsEmpty()); |
205 DCHECK(tz->IsEmpty()); | 202 DCHECK(tz->IsEmpty()); |
206 | 203 |
207 // Parse mandatory date string: [('-'|'+')yy]yyyy[':'MM[':'DD]] | 204 // Parse mandatory date string: [('-'|'+')yy]yyyy[':'MM[':'DD]] |
208 if (scanner->Peek().IsAsciiSign()) { | 205 if (scanner->Peek().IsAsciiSign()) { |
(...skipping 17 matching lines...) Expand all Loading... |
226 if (scanner->SkipSymbol('-')) { | 223 if (scanner->SkipSymbol('-')) { |
227 if (!scanner->Peek().IsFixedLengthNumber(2) || | 224 if (!scanner->Peek().IsFixedLengthNumber(2) || |
228 !DayComposer::IsDay(scanner->Peek().number())) return scanner->Next(); | 225 !DayComposer::IsDay(scanner->Peek().number())) return scanner->Next(); |
229 day->Add(scanner->Next().number()); | 226 day->Add(scanner->Next().number()); |
230 } | 227 } |
231 } | 228 } |
232 // Check for optional time string: 'T'HH':'mm[':'ss['.'sss]]Z | 229 // Check for optional time string: 'T'HH':'mm[':'ss['.'sss]]Z |
233 if (!scanner->Peek().IsKeywordType(TIME_SEPARATOR)) { | 230 if (!scanner->Peek().IsKeywordType(TIME_SEPARATOR)) { |
234 if (!scanner->Peek().IsEndOfInput()) return scanner->Next(); | 231 if (!scanner->Peek().IsEndOfInput()) return scanner->Next(); |
235 } else { | 232 } else { |
236 // ES5 Date Time String time part is present. | 233 // ES6 Date Time String time part is present. |
237 scanner->Next(); | 234 scanner->Next(); |
238 if (!scanner->Peek().IsFixedLengthNumber(2) || | 235 if (!scanner->Peek().IsFixedLengthNumber(2) || |
239 !Between(scanner->Peek().number(), 0, 24)) { | 236 !Between(scanner->Peek().number(), 0, 24)) { |
240 return DateToken::Invalid(); | 237 return DateToken::Invalid(); |
241 } | 238 } |
242 // Allow 24:00[:00[.000]], but no other time starting with 24. | 239 // Allow 24:00[:00[.000]], but no other time starting with 24. |
243 bool hour_is_24 = (scanner->Peek().number() == 24); | 240 bool hour_is_24 = (scanner->Peek().number() == 24); |
244 time->Add(scanner->Next().number()); | 241 time->Add(scanner->Next().number()); |
245 if (!scanner->SkipSymbol(':')) return DateToken::Invalid(); | 242 if (!scanner->SkipSymbol(':')) return DateToken::Invalid(); |
246 if (!scanner->Peek().IsFixedLengthNumber(2) || | 243 if (!scanner->Peek().IsFixedLengthNumber(2) || |
(...skipping 45 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
292 if (!scanner->SkipSymbol(':')) return DateToken::Invalid(); | 289 if (!scanner->SkipSymbol(':')) return DateToken::Invalid(); |
293 if (!scanner->Peek().IsFixedLengthNumber(2) || | 290 if (!scanner->Peek().IsFixedLengthNumber(2) || |
294 !TimeComposer::IsMinute(scanner->Peek().number())) { | 291 !TimeComposer::IsMinute(scanner->Peek().number())) { |
295 return DateToken::Invalid(); | 292 return DateToken::Invalid(); |
296 } | 293 } |
297 tz->SetAbsoluteMinute(scanner->Next().number()); | 294 tz->SetAbsoluteMinute(scanner->Next().number()); |
298 } | 295 } |
299 } | 296 } |
300 if (!scanner->Peek().IsEndOfInput()) return DateToken::Invalid(); | 297 if (!scanner->Peek().IsEndOfInput()) return DateToken::Invalid(); |
301 } | 298 } |
302 // Successfully parsed ES5 Date Time String. Default to UTC if no TZ given. | 299 // Successfully parsed ES6 Date Time String. |
303 if (tz->IsEmpty()) tz->Set(0); | |
304 day->set_iso_date(); | 300 day->set_iso_date(); |
305 return DateToken::EndOfInput(); | 301 return DateToken::EndOfInput(); |
306 } | 302 } |
307 | 303 |
308 | 304 |
309 } } // namespace v8::internal | 305 } } // namespace v8::internal |
310 | 306 |
311 #endif // V8_DATEPARSER_INL_H_ | 307 #endif // V8_DATEPARSER_INL_H_ |
OLD | NEW |