/rust/registry/src/index.crates.io-1949cf8c6b5b557f/time-0.3.44/src/parsing/parsable.rs
Line | Count | Source |
1 | | //! A trait that can be used to parse an item from an input. |
2 | | |
3 | | use core::ops::Deref; |
4 | | |
5 | | use num_conv::prelude::*; |
6 | | |
7 | | use crate::error::TryFromParsed; |
8 | | use crate::format_description::well_known::iso8601::EncodedConfig; |
9 | | use crate::format_description::well_known::{Iso8601, Rfc2822, Rfc3339}; |
10 | | use crate::format_description::BorrowedFormatItem; |
11 | | #[cfg(feature = "alloc")] |
12 | | use crate::format_description::OwnedFormatItem; |
13 | | use crate::internal_macros::bug; |
14 | | use crate::parsing::{Parsed, ParsedItem}; |
15 | | use crate::{error, Date, Month, OffsetDateTime, Time, UtcOffset, Weekday}; |
16 | | |
17 | | /// A type that can be parsed. |
18 | | #[cfg_attr(docsrs, doc(notable_trait))] |
19 | | #[doc(alias = "Parseable")] |
20 | | pub trait Parsable: sealed::Sealed {} |
21 | | impl Parsable for BorrowedFormatItem<'_> {} |
22 | | impl Parsable for [BorrowedFormatItem<'_>] {} |
23 | | #[cfg(feature = "alloc")] |
24 | | impl Parsable for OwnedFormatItem {} |
25 | | #[cfg(feature = "alloc")] |
26 | | impl Parsable for [OwnedFormatItem] {} |
27 | | impl Parsable for Rfc2822 {} |
28 | | impl Parsable for Rfc3339 {} |
29 | | impl<const CONFIG: EncodedConfig> Parsable for Iso8601<CONFIG> {} |
30 | | impl<T: Deref> Parsable for T where T::Target: Parsable {} |
31 | | |
32 | | /// Seal the trait to prevent downstream users from implementing it, while still allowing it to |
33 | | /// exist in generic bounds. |
34 | | mod sealed { |
35 | | use super::*; |
36 | | use crate::{PrimitiveDateTime, UtcDateTime}; |
37 | | |
38 | | /// Parse the item using a format description and an input. |
39 | | pub trait Sealed { |
40 | | /// Parse the item into the provided [`Parsed`] struct. |
41 | | /// |
42 | | /// This method can be used to parse a single component without parsing the full value. |
43 | | fn parse_into<'a>( |
44 | | &self, |
45 | | input: &'a [u8], |
46 | | parsed: &mut Parsed, |
47 | | ) -> Result<&'a [u8], error::Parse>; |
48 | | |
49 | | /// Parse the item into a new [`Parsed`] struct. |
50 | | /// |
51 | | /// This method can only be used to parse a complete value of a type. If any characters |
52 | | /// remain after parsing, an error will be returned. |
53 | | #[inline] |
54 | 0 | fn parse(&self, input: &[u8]) -> Result<Parsed, error::Parse> { |
55 | 0 | let mut parsed = Parsed::new(); |
56 | 0 | if self.parse_into(input, &mut parsed)?.is_empty() { |
57 | 0 | Ok(parsed) |
58 | | } else { |
59 | 0 | Err(error::Parse::ParseFromDescription( |
60 | 0 | error::ParseFromDescription::UnexpectedTrailingCharacters, |
61 | 0 | )) |
62 | | } |
63 | 0 | } |
64 | | |
65 | | /// Parse a [`Date`] from the format description. |
66 | | #[inline] |
67 | 0 | fn parse_date(&self, input: &[u8]) -> Result<Date, error::Parse> { |
68 | 0 | Ok(self.parse(input)?.try_into()?) |
69 | 0 | } |
70 | | |
71 | | /// Parse a [`Time`] from the format description. |
72 | | #[inline] |
73 | 0 | fn parse_time(&self, input: &[u8]) -> Result<Time, error::Parse> { |
74 | 0 | Ok(self.parse(input)?.try_into()?) |
75 | 0 | } |
76 | | |
77 | | /// Parse a [`UtcOffset`] from the format description. |
78 | | #[inline] |
79 | 0 | fn parse_offset(&self, input: &[u8]) -> Result<UtcOffset, error::Parse> { |
80 | 0 | Ok(self.parse(input)?.try_into()?) |
81 | 0 | } |
82 | | |
83 | | /// Parse a [`PrimitiveDateTime`] from the format description. |
84 | | #[inline] |
85 | 0 | fn parse_primitive_date_time( |
86 | 0 | &self, |
87 | 0 | input: &[u8], |
88 | 0 | ) -> Result<PrimitiveDateTime, error::Parse> { |
89 | 0 | Ok(self.parse(input)?.try_into()?) |
90 | 0 | } |
91 | | |
92 | | /// Parse a [`UtcDateTime`] from the format description. |
93 | | #[inline] |
94 | 0 | fn parse_utc_date_time(&self, input: &[u8]) -> Result<UtcDateTime, error::Parse> { |
95 | 0 | Ok(self.parse(input)?.try_into()?) |
96 | 0 | } |
97 | | |
98 | | /// Parse a [`OffsetDateTime`] from the format description. |
99 | | #[inline] |
100 | 0 | fn parse_offset_date_time(&self, input: &[u8]) -> Result<OffsetDateTime, error::Parse> { |
101 | 0 | Ok(self.parse(input)?.try_into()?) |
102 | 0 | } |
103 | | } |
104 | | } |
105 | | |
106 | | impl sealed::Sealed for BorrowedFormatItem<'_> { |
107 | | #[inline] |
108 | 0 | fn parse_into<'a>( |
109 | 0 | &self, |
110 | 0 | input: &'a [u8], |
111 | 0 | parsed: &mut Parsed, |
112 | 0 | ) -> Result<&'a [u8], error::Parse> { |
113 | 0 | Ok(parsed.parse_item(input, self)?) |
114 | 0 | } |
115 | | } |
116 | | |
117 | | impl sealed::Sealed for [BorrowedFormatItem<'_>] { |
118 | | #[inline] |
119 | 0 | fn parse_into<'a>( |
120 | 0 | &self, |
121 | 0 | input: &'a [u8], |
122 | 0 | parsed: &mut Parsed, |
123 | 0 | ) -> Result<&'a [u8], error::Parse> { |
124 | 0 | Ok(parsed.parse_items(input, self)?) |
125 | 0 | } |
126 | | } |
127 | | |
128 | | #[cfg(feature = "alloc")] |
129 | | impl sealed::Sealed for OwnedFormatItem { |
130 | | #[inline] |
131 | 0 | fn parse_into<'a>( |
132 | 0 | &self, |
133 | 0 | input: &'a [u8], |
134 | 0 | parsed: &mut Parsed, |
135 | 0 | ) -> Result<&'a [u8], error::Parse> { |
136 | 0 | Ok(parsed.parse_item(input, self)?) |
137 | 0 | } |
138 | | } |
139 | | |
140 | | #[cfg(feature = "alloc")] |
141 | | impl sealed::Sealed for [OwnedFormatItem] { |
142 | | #[inline] |
143 | 0 | fn parse_into<'a>( |
144 | 0 | &self, |
145 | 0 | input: &'a [u8], |
146 | 0 | parsed: &mut Parsed, |
147 | 0 | ) -> Result<&'a [u8], error::Parse> { |
148 | 0 | Ok(parsed.parse_items(input, self)?) |
149 | 0 | } |
150 | | } |
151 | | |
152 | | impl<T> sealed::Sealed for T |
153 | | where |
154 | | T: Deref<Target: sealed::Sealed>, |
155 | | { |
156 | | #[inline] |
157 | 0 | fn parse_into<'a>( |
158 | 0 | &self, |
159 | 0 | input: &'a [u8], |
160 | 0 | parsed: &mut Parsed, |
161 | 0 | ) -> Result<&'a [u8], error::Parse> { |
162 | 0 | self.deref().parse_into(input, parsed) |
163 | 0 | } |
164 | | } |
165 | | |
166 | | impl sealed::Sealed for Rfc2822 { |
167 | 0 | fn parse_into<'a>( |
168 | 0 | &self, |
169 | 0 | input: &'a [u8], |
170 | 0 | parsed: &mut Parsed, |
171 | 0 | ) -> Result<&'a [u8], error::Parse> { |
172 | | use crate::error::ParseFromDescription::{InvalidComponent, InvalidLiteral}; |
173 | | use crate::parsing::combinator::rfc::rfc2822::{cfws, fws}; |
174 | | use crate::parsing::combinator::{ |
175 | | ascii_char, exactly_n_digits, first_match, n_to_m_digits, opt, sign, |
176 | | }; |
177 | | |
178 | 0 | let colon = ascii_char::<b':'>; |
179 | 0 | let comma = ascii_char::<b','>; |
180 | | |
181 | 0 | let input = opt(cfws)(input).into_inner(); |
182 | 0 | let weekday = first_match( |
183 | 0 | [ |
184 | 0 | (b"Mon".as_slice(), Weekday::Monday), |
185 | 0 | (b"Tue".as_slice(), Weekday::Tuesday), |
186 | 0 | (b"Wed".as_slice(), Weekday::Wednesday), |
187 | 0 | (b"Thu".as_slice(), Weekday::Thursday), |
188 | 0 | (b"Fri".as_slice(), Weekday::Friday), |
189 | 0 | (b"Sat".as_slice(), Weekday::Saturday), |
190 | 0 | (b"Sun".as_slice(), Weekday::Sunday), |
191 | 0 | ], |
192 | 0 | false, |
193 | 0 | )(input); |
194 | 0 | let input = if let Some(item) = weekday { |
195 | 0 | let input = item |
196 | 0 | .consume_value(|value| parsed.set_weekday(value)) |
197 | 0 | .ok_or(InvalidComponent("weekday"))?; |
198 | 0 | let input = comma(input).ok_or(InvalidLiteral)?.into_inner(); |
199 | 0 | opt(cfws)(input).into_inner() |
200 | | } else { |
201 | 0 | input |
202 | | }; |
203 | 0 | let input = n_to_m_digits::<1, 2, _>(input) |
204 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_day(value))) |
205 | 0 | .ok_or(InvalidComponent("day"))?; |
206 | 0 | let input = cfws(input).ok_or(InvalidLiteral)?.into_inner(); |
207 | 0 | let input = first_match( |
208 | 0 | [ |
209 | 0 | (b"Jan".as_slice(), Month::January), |
210 | 0 | (b"Feb".as_slice(), Month::February), |
211 | 0 | (b"Mar".as_slice(), Month::March), |
212 | 0 | (b"Apr".as_slice(), Month::April), |
213 | 0 | (b"May".as_slice(), Month::May), |
214 | 0 | (b"Jun".as_slice(), Month::June), |
215 | 0 | (b"Jul".as_slice(), Month::July), |
216 | 0 | (b"Aug".as_slice(), Month::August), |
217 | 0 | (b"Sep".as_slice(), Month::September), |
218 | 0 | (b"Oct".as_slice(), Month::October), |
219 | 0 | (b"Nov".as_slice(), Month::November), |
220 | 0 | (b"Dec".as_slice(), Month::December), |
221 | 0 | ], |
222 | 0 | false, |
223 | 0 | )(input) |
224 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_month(value))) |
225 | 0 | .ok_or(InvalidComponent("month"))?; |
226 | 0 | let input = cfws(input).ok_or(InvalidLiteral)?.into_inner(); |
227 | 0 | let input = match exactly_n_digits::<4, u32>(input) { |
228 | 0 | Some(item) => { |
229 | 0 | let input = item |
230 | 0 | .flat_map(|year| if year >= 1900 { Some(year) } else { None }) |
231 | 0 | .and_then(|item| { |
232 | 0 | item.consume_value(|value| parsed.set_year(value.cast_signed())) |
233 | 0 | }) |
234 | 0 | .ok_or(InvalidComponent("year"))?; |
235 | 0 | fws(input).ok_or(InvalidLiteral)?.into_inner() |
236 | | } |
237 | | None => { |
238 | 0 | let input = exactly_n_digits::<2, u32>(input) |
239 | 0 | .and_then(|item| { |
240 | 0 | item.map(|year| if year < 50 { year + 2000 } else { year + 1900 }) |
241 | 0 | .map(|year| year.cast_signed()) |
242 | 0 | .consume_value(|value| parsed.set_year(value)) |
243 | 0 | }) |
244 | 0 | .ok_or(InvalidComponent("year"))?; |
245 | 0 | cfws(input).ok_or(InvalidLiteral)?.into_inner() |
246 | | } |
247 | | }; |
248 | | |
249 | 0 | let input = exactly_n_digits::<2, _>(input) |
250 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_hour_24(value))) |
251 | 0 | .ok_or(InvalidComponent("hour"))?; |
252 | 0 | let input = opt(cfws)(input).into_inner(); |
253 | 0 | let input = colon(input).ok_or(InvalidLiteral)?.into_inner(); |
254 | 0 | let input = opt(cfws)(input).into_inner(); |
255 | 0 | let input = exactly_n_digits::<2, _>(input) |
256 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_minute(value))) |
257 | 0 | .ok_or(InvalidComponent("minute"))?; |
258 | | |
259 | 0 | let input = if let Some(input) = colon(opt(cfws)(input).into_inner()) { |
260 | 0 | let input = input.into_inner(); // discard the colon |
261 | 0 | let input = opt(cfws)(input).into_inner(); |
262 | 0 | let input = exactly_n_digits::<2, _>(input) |
263 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_second(value))) |
264 | 0 | .ok_or(InvalidComponent("second"))?; |
265 | 0 | cfws(input).ok_or(InvalidLiteral)?.into_inner() |
266 | | } else { |
267 | 0 | cfws(input).ok_or(InvalidLiteral)?.into_inner() |
268 | | }; |
269 | | |
270 | | // The RFC explicitly allows leap seconds. |
271 | 0 | parsed.leap_second_allowed = true; |
272 | | |
273 | | #[expect( |
274 | | clippy::unnecessary_lazy_evaluations, |
275 | | reason = "rust-lang/rust-clippy#8522" |
276 | | )] |
277 | 0 | let zone_literal = first_match( |
278 | 0 | [ |
279 | 0 | (b"UT".as_slice(), 0), |
280 | 0 | (b"GMT".as_slice(), 0), |
281 | 0 | (b"EST".as_slice(), -5), |
282 | 0 | (b"EDT".as_slice(), -4), |
283 | 0 | (b"CST".as_slice(), -6), |
284 | 0 | (b"CDT".as_slice(), -5), |
285 | 0 | (b"MST".as_slice(), -7), |
286 | 0 | (b"MDT".as_slice(), -6), |
287 | 0 | (b"PST".as_slice(), -8), |
288 | 0 | (b"PDT".as_slice(), -7), |
289 | 0 | ], |
290 | 0 | false, |
291 | 0 | )(input) |
292 | 0 | .or_else(|| match input { |
293 | 0 | [b'a'..=b'i' | b'k'..=b'z' | b'A'..=b'I' | b'K'..=b'Z', rest @ ..] => { |
294 | 0 | Some(ParsedItem(rest, 0)) |
295 | | } |
296 | 0 | _ => None, |
297 | 0 | }); |
298 | 0 | if let Some(zone_literal) = zone_literal { |
299 | 0 | let input = zone_literal |
300 | 0 | .consume_value(|value| parsed.set_offset_hour(value)) |
301 | 0 | .ok_or(InvalidComponent("offset hour"))?; |
302 | 0 | parsed |
303 | 0 | .set_offset_minute_signed(0) |
304 | 0 | .ok_or(InvalidComponent("offset minute"))?; |
305 | 0 | parsed |
306 | 0 | .set_offset_second_signed(0) |
307 | 0 | .ok_or(InvalidComponent("offset second"))?; |
308 | 0 | return Ok(input); |
309 | 0 | } |
310 | | |
311 | 0 | let ParsedItem(input, offset_sign) = sign(input).ok_or(InvalidComponent("offset hour"))?; |
312 | 0 | let input = exactly_n_digits::<2, u8>(input) |
313 | 0 | .and_then(|item| { |
314 | 0 | item.map(|offset_hour| { |
315 | 0 | if offset_sign == b'-' { |
316 | 0 | -offset_hour.cast_signed() |
317 | | } else { |
318 | 0 | offset_hour.cast_signed() |
319 | | } |
320 | 0 | }) |
321 | 0 | .consume_value(|value| parsed.set_offset_hour(value)) |
322 | 0 | }) |
323 | 0 | .ok_or(InvalidComponent("offset hour"))?; |
324 | 0 | let input = exactly_n_digits::<2, u8>(input) |
325 | 0 | .and_then(|item| { |
326 | 0 | item.consume_value(|value| parsed.set_offset_minute_signed(value.cast_signed())) |
327 | 0 | }) |
328 | 0 | .ok_or(InvalidComponent("offset minute"))?; |
329 | | |
330 | 0 | let input = opt(cfws)(input).into_inner(); |
331 | | |
332 | 0 | Ok(input) |
333 | 0 | } |
334 | | |
335 | 0 | fn parse_offset_date_time(&self, input: &[u8]) -> Result<OffsetDateTime, error::Parse> { |
336 | | use crate::error::ParseFromDescription::{InvalidComponent, InvalidLiteral}; |
337 | | use crate::parsing::combinator::rfc::rfc2822::{cfws, fws}; |
338 | | use crate::parsing::combinator::{ |
339 | | ascii_char, exactly_n_digits, first_match, n_to_m_digits, opt, sign, |
340 | | }; |
341 | | |
342 | 0 | let colon = ascii_char::<b':'>; |
343 | 0 | let comma = ascii_char::<b','>; |
344 | | |
345 | 0 | let input = opt(cfws)(input).into_inner(); |
346 | | // This parses the weekday, but we don't actually use the value anywhere. Because of this, |
347 | | // just return `()` to avoid unnecessary generated code. |
348 | 0 | let weekday = first_match( |
349 | 0 | [ |
350 | 0 | (b"Mon".as_slice(), ()), |
351 | 0 | (b"Tue".as_slice(), ()), |
352 | 0 | (b"Wed".as_slice(), ()), |
353 | 0 | (b"Thu".as_slice(), ()), |
354 | 0 | (b"Fri".as_slice(), ()), |
355 | 0 | (b"Sat".as_slice(), ()), |
356 | 0 | (b"Sun".as_slice(), ()), |
357 | 0 | ], |
358 | 0 | false, |
359 | 0 | )(input); |
360 | 0 | let input = if let Some(item) = weekday { |
361 | 0 | let input = item.into_inner(); |
362 | 0 | let input = comma(input).ok_or(InvalidLiteral)?.into_inner(); |
363 | 0 | opt(cfws)(input).into_inner() |
364 | | } else { |
365 | 0 | input |
366 | | }; |
367 | 0 | let ParsedItem(input, day) = |
368 | 0 | n_to_m_digits::<1, 2, _>(input).ok_or(InvalidComponent("day"))?; |
369 | 0 | let input = cfws(input).ok_or(InvalidLiteral)?.into_inner(); |
370 | 0 | let ParsedItem(input, month) = first_match( |
371 | 0 | [ |
372 | 0 | (b"Jan".as_slice(), Month::January), |
373 | 0 | (b"Feb".as_slice(), Month::February), |
374 | 0 | (b"Mar".as_slice(), Month::March), |
375 | 0 | (b"Apr".as_slice(), Month::April), |
376 | 0 | (b"May".as_slice(), Month::May), |
377 | 0 | (b"Jun".as_slice(), Month::June), |
378 | 0 | (b"Jul".as_slice(), Month::July), |
379 | 0 | (b"Aug".as_slice(), Month::August), |
380 | 0 | (b"Sep".as_slice(), Month::September), |
381 | 0 | (b"Oct".as_slice(), Month::October), |
382 | 0 | (b"Nov".as_slice(), Month::November), |
383 | 0 | (b"Dec".as_slice(), Month::December), |
384 | 0 | ], |
385 | 0 | false, |
386 | 0 | )(input) |
387 | 0 | .ok_or(InvalidComponent("month"))?; |
388 | 0 | let input = cfws(input).ok_or(InvalidLiteral)?.into_inner(); |
389 | 0 | let (input, year) = match exactly_n_digits::<4, u32>(input) { |
390 | 0 | Some(item) => { |
391 | 0 | let ParsedItem(input, year) = item |
392 | 0 | .flat_map(|year| if year >= 1900 { Some(year) } else { None }) |
393 | 0 | .ok_or(InvalidComponent("year"))?; |
394 | 0 | let input = fws(input).ok_or(InvalidLiteral)?.into_inner(); |
395 | 0 | (input, year) |
396 | | } |
397 | | None => { |
398 | 0 | let ParsedItem(input, year) = exactly_n_digits::<2, u32>(input) |
399 | 0 | .map(|item| item.map(|year| if year < 50 { year + 2000 } else { year + 1900 })) |
400 | 0 | .ok_or(InvalidComponent("year"))?; |
401 | 0 | let input = cfws(input).ok_or(InvalidLiteral)?.into_inner(); |
402 | 0 | (input, year) |
403 | | } |
404 | | }; |
405 | | |
406 | 0 | let ParsedItem(input, hour) = |
407 | 0 | exactly_n_digits::<2, _>(input).ok_or(InvalidComponent("hour"))?; |
408 | 0 | let input = opt(cfws)(input).into_inner(); |
409 | 0 | let input = colon(input).ok_or(InvalidLiteral)?.into_inner(); |
410 | 0 | let input = opt(cfws)(input).into_inner(); |
411 | 0 | let ParsedItem(input, minute) = |
412 | 0 | exactly_n_digits::<2, _>(input).ok_or(InvalidComponent("minute"))?; |
413 | | |
414 | 0 | let (input, mut second) = if let Some(input) = colon(opt(cfws)(input).into_inner()) { |
415 | 0 | let input = input.into_inner(); // discard the colon |
416 | 0 | let input = opt(cfws)(input).into_inner(); |
417 | 0 | let ParsedItem(input, second) = |
418 | 0 | exactly_n_digits::<2, _>(input).ok_or(InvalidComponent("second"))?; |
419 | 0 | let input = cfws(input).ok_or(InvalidLiteral)?.into_inner(); |
420 | 0 | (input, second) |
421 | | } else { |
422 | 0 | (cfws(input).ok_or(InvalidLiteral)?.into_inner(), 0) |
423 | | }; |
424 | | |
425 | | #[expect( |
426 | | clippy::unnecessary_lazy_evaluations, |
427 | | reason = "rust-lang/rust-clippy#8522" |
428 | | )] |
429 | 0 | let zone_literal = first_match( |
430 | 0 | [ |
431 | 0 | (b"UT".as_slice(), 0), |
432 | 0 | (b"GMT".as_slice(), 0), |
433 | 0 | (b"EST".as_slice(), -5), |
434 | 0 | (b"EDT".as_slice(), -4), |
435 | 0 | (b"CST".as_slice(), -6), |
436 | 0 | (b"CDT".as_slice(), -5), |
437 | 0 | (b"MST".as_slice(), -7), |
438 | 0 | (b"MDT".as_slice(), -6), |
439 | 0 | (b"PST".as_slice(), -8), |
440 | 0 | (b"PDT".as_slice(), -7), |
441 | 0 | ], |
442 | 0 | false, |
443 | 0 | )(input) |
444 | 0 | .or_else(|| match input { |
445 | 0 | [b'a'..=b'i' | b'k'..=b'z' | b'A'..=b'I' | b'K'..=b'Z', rest @ ..] => { |
446 | 0 | Some(ParsedItem(rest, 0)) |
447 | | } |
448 | 0 | _ => None, |
449 | 0 | }); |
450 | | |
451 | 0 | let (input, offset_hour, offset_minute) = if let Some(zone_literal) = zone_literal { |
452 | 0 | let ParsedItem(input, offset_hour) = zone_literal; |
453 | 0 | (input, offset_hour, 0) |
454 | | } else { |
455 | 0 | let ParsedItem(input, offset_sign) = |
456 | 0 | sign(input).ok_or(InvalidComponent("offset hour"))?; |
457 | 0 | let ParsedItem(input, offset_hour) = exactly_n_digits::<2, u8>(input) |
458 | 0 | .map(|item| { |
459 | 0 | item.map(|offset_hour| { |
460 | 0 | if offset_sign == b'-' { |
461 | 0 | -offset_hour.cast_signed() |
462 | | } else { |
463 | 0 | offset_hour.cast_signed() |
464 | | } |
465 | 0 | }) |
466 | 0 | }) |
467 | 0 | .ok_or(InvalidComponent("offset hour"))?; |
468 | 0 | let ParsedItem(input, offset_minute) = |
469 | 0 | exactly_n_digits::<2, u8>(input).ok_or(InvalidComponent("offset minute"))?; |
470 | 0 | (input, offset_hour, offset_minute.cast_signed()) |
471 | | }; |
472 | | |
473 | 0 | let input = opt(cfws)(input).into_inner(); |
474 | | |
475 | 0 | if !input.is_empty() { |
476 | 0 | return Err(error::Parse::ParseFromDescription( |
477 | 0 | error::ParseFromDescription::UnexpectedTrailingCharacters, |
478 | 0 | )); |
479 | 0 | } |
480 | | |
481 | 0 | let mut nanosecond = 0; |
482 | 0 | let leap_second_input = if second == 60 { |
483 | 0 | second = 59; |
484 | 0 | nanosecond = 999_999_999; |
485 | 0 | true |
486 | | } else { |
487 | 0 | false |
488 | | }; |
489 | | |
490 | 0 | let dt = (|| { |
491 | 0 | let date = Date::from_calendar_date(year.cast_signed(), month, day)?; |
492 | 0 | let time = Time::from_hms_nano(hour, minute, second, nanosecond)?; |
493 | 0 | let offset = UtcOffset::from_hms(offset_hour, offset_minute, 0)?; |
494 | 0 | Ok(OffsetDateTime::new_in_offset(date, time, offset)) |
495 | | })() |
496 | 0 | .map_err(TryFromParsed::ComponentRange)?; |
497 | | |
498 | 0 | if leap_second_input && !dt.is_valid_leap_second_stand_in() { |
499 | 0 | return Err(error::Parse::TryFromParsed(TryFromParsed::ComponentRange( |
500 | 0 | error::ComponentRange { |
501 | 0 | name: "second", |
502 | 0 | minimum: 0, |
503 | 0 | maximum: 59, |
504 | 0 | value: 60, |
505 | 0 | conditional_message: Some("because leap seconds are not supported"), |
506 | 0 | }, |
507 | 0 | ))); |
508 | 0 | } |
509 | | |
510 | 0 | Ok(dt) |
511 | 0 | } |
512 | | } |
513 | | |
514 | | impl sealed::Sealed for Rfc3339 { |
515 | 0 | fn parse_into<'a>( |
516 | 0 | &self, |
517 | 0 | input: &'a [u8], |
518 | 0 | parsed: &mut Parsed, |
519 | 0 | ) -> Result<&'a [u8], error::Parse> { |
520 | | use crate::error::ParseFromDescription::{InvalidComponent, InvalidLiteral}; |
521 | | use crate::parsing::combinator::{ |
522 | | any_digit, ascii_char, ascii_char_ignore_case, exactly_n_digits, sign, |
523 | | }; |
524 | | |
525 | 0 | let dash = ascii_char::<b'-'>; |
526 | 0 | let colon = ascii_char::<b':'>; |
527 | | |
528 | 0 | let input = exactly_n_digits::<4, u32>(input) |
529 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_year(value.cast_signed()))) |
530 | 0 | .ok_or(InvalidComponent("year"))?; |
531 | 0 | let input = dash(input).ok_or(InvalidLiteral)?.into_inner(); |
532 | 0 | let input = exactly_n_digits::<2, _>(input) |
533 | 0 | .and_then(|item| item.flat_map(|value| Month::from_number(value).ok())) |
534 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_month(value))) |
535 | 0 | .ok_or(InvalidComponent("month"))?; |
536 | 0 | let input = dash(input).ok_or(InvalidLiteral)?.into_inner(); |
537 | 0 | let input = exactly_n_digits::<2, _>(input) |
538 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_day(value))) |
539 | 0 | .ok_or(InvalidComponent("day"))?; |
540 | | |
541 | | // RFC3339 allows any separator, not just `T`, not just `space`. |
542 | | // cf. Section 5.6: Internet Date/Time Format: |
543 | | // NOTE: ISO 8601 defines date and time separated by "T". |
544 | | // Applications using this syntax may choose, for the sake of |
545 | | // readability, to specify a full-date and full-time separated by |
546 | | // (say) a space character. |
547 | | // Specifically, rusqlite uses space separators. |
548 | 0 | let input = input.get(1..).ok_or(InvalidComponent("separator"))?; |
549 | | |
550 | 0 | let input = exactly_n_digits::<2, _>(input) |
551 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_hour_24(value))) |
552 | 0 | .ok_or(InvalidComponent("hour"))?; |
553 | 0 | let input = colon(input).ok_or(InvalidLiteral)?.into_inner(); |
554 | 0 | let input = exactly_n_digits::<2, _>(input) |
555 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_minute(value))) |
556 | 0 | .ok_or(InvalidComponent("minute"))?; |
557 | 0 | let input = colon(input).ok_or(InvalidLiteral)?.into_inner(); |
558 | 0 | let input = exactly_n_digits::<2, _>(input) |
559 | 0 | .and_then(|item| item.consume_value(|value| parsed.set_second(value))) |
560 | 0 | .ok_or(InvalidComponent("second"))?; |
561 | 0 | let input = if let Some(ParsedItem(input, ())) = ascii_char::<b'.'>(input) { |
562 | 0 | let ParsedItem(mut input, mut value) = any_digit(input) |
563 | 0 | .ok_or(InvalidComponent("subsecond"))? |
564 | 0 | .map(|v| (v - b'0').extend::<u32>() * 100_000_000); |
565 | | |
566 | 0 | let mut multiplier = 10_000_000; |
567 | 0 | while let Some(ParsedItem(new_input, digit)) = any_digit(input) { |
568 | 0 | value += (digit - b'0').extend::<u32>() * multiplier; |
569 | 0 | input = new_input; |
570 | 0 | multiplier /= 10; |
571 | 0 | } |
572 | | |
573 | 0 | parsed |
574 | 0 | .set_subsecond(value) |
575 | 0 | .ok_or(InvalidComponent("subsecond"))?; |
576 | 0 | input |
577 | | } else { |
578 | 0 | input |
579 | | }; |
580 | | |
581 | | // The RFC explicitly allows leap seconds. |
582 | 0 | parsed.leap_second_allowed = true; |
583 | | |
584 | 0 | if let Some(ParsedItem(input, ())) = ascii_char_ignore_case::<b'Z'>(input) { |
585 | 0 | parsed |
586 | 0 | .set_offset_hour(0) |
587 | 0 | .ok_or(InvalidComponent("offset hour"))?; |
588 | 0 | parsed |
589 | 0 | .set_offset_minute_signed(0) |
590 | 0 | .ok_or(InvalidComponent("offset minute"))?; |
591 | 0 | parsed |
592 | 0 | .set_offset_second_signed(0) |
593 | 0 | .ok_or(InvalidComponent("offset second"))?; |
594 | 0 | return Ok(input); |
595 | 0 | } |
596 | | |
597 | 0 | let ParsedItem(input, offset_sign) = sign(input).ok_or(InvalidComponent("offset hour"))?; |
598 | 0 | let input = exactly_n_digits::<2, u8>(input) |
599 | 0 | .and_then(|item| { |
600 | 0 | item.filter(|&offset_hour| offset_hour <= 23)? |
601 | 0 | .map(|offset_hour| { |
602 | 0 | if offset_sign == b'-' { |
603 | 0 | -offset_hour.cast_signed() |
604 | | } else { |
605 | 0 | offset_hour.cast_signed() |
606 | | } |
607 | 0 | }) |
608 | 0 | .consume_value(|value| parsed.set_offset_hour(value)) |
609 | 0 | }) |
610 | 0 | .ok_or(InvalidComponent("offset hour"))?; |
611 | 0 | let input = colon(input).ok_or(InvalidLiteral)?.into_inner(); |
612 | 0 | let input = exactly_n_digits::<2, u8>(input) |
613 | 0 | .and_then(|item| { |
614 | 0 | item.map(|offset_minute| { |
615 | 0 | if offset_sign == b'-' { |
616 | 0 | -offset_minute.cast_signed() |
617 | | } else { |
618 | 0 | offset_minute.cast_signed() |
619 | | } |
620 | 0 | }) |
621 | 0 | .consume_value(|value| parsed.set_offset_minute_signed(value)) |
622 | 0 | }) |
623 | 0 | .ok_or(InvalidComponent("offset minute"))?; |
624 | | |
625 | 0 | Ok(input) |
626 | 0 | } |
627 | | |
628 | 23.3k | fn parse_offset_date_time(&self, input: &[u8]) -> Result<OffsetDateTime, error::Parse> { |
629 | | use crate::error::ParseFromDescription::{InvalidComponent, InvalidLiteral}; |
630 | | use crate::parsing::combinator::{ |
631 | | any_digit, ascii_char, ascii_char_ignore_case, exactly_n_digits, sign, |
632 | | }; |
633 | | |
634 | 23.3k | let dash = ascii_char::<b'-'>; |
635 | 23.3k | let colon = ascii_char::<b':'>; |
636 | | |
637 | 23.1k | let ParsedItem(input, year) = |
638 | 23.3k | exactly_n_digits::<4, u32>(input).ok_or(InvalidComponent("year"))?; |
639 | 23.1k | let input = dash(input).ok_or(InvalidLiteral)?.into_inner(); |
640 | 23.1k | let ParsedItem(input, month) = |
641 | 23.1k | exactly_n_digits::<2, _>(input).ok_or(InvalidComponent("month"))?; |
642 | 23.1k | let input = dash(input).ok_or(InvalidLiteral)?.into_inner(); |
643 | 23.0k | let ParsedItem(input, day) = |
644 | 23.1k | exactly_n_digits::<2, _>(input).ok_or(InvalidComponent("day"))?; |
645 | | |
646 | | // RFC3339 allows any separator, not just `T`, not just `space`. |
647 | | // cf. Section 5.6: Internet Date/Time Format: |
648 | | // NOTE: ISO 8601 defines date and time separated by "T". |
649 | | // Applications using this syntax may choose, for the sake of |
650 | | // readability, to specify a full-date and full-time separated by |
651 | | // (say) a space character. |
652 | | // Specifically, rusqlite uses space separators. |
653 | 23.0k | let input = input.get(1..).ok_or(InvalidComponent("separator"))?; |
654 | | |
655 | 23.0k | let ParsedItem(input, hour) = |
656 | 23.0k | exactly_n_digits::<2, _>(input).ok_or(InvalidComponent("hour"))?; |
657 | 23.0k | let input = colon(input).ok_or(InvalidLiteral)?.into_inner(); |
658 | 22.9k | let ParsedItem(input, minute) = |
659 | 22.9k | exactly_n_digits::<2, _>(input).ok_or(InvalidComponent("minute"))?; |
660 | 22.9k | let input = colon(input).ok_or(InvalidLiteral)?.into_inner(); |
661 | 22.9k | let ParsedItem(input, mut second) = |
662 | 22.9k | exactly_n_digits::<2, _>(input).ok_or(InvalidComponent("second"))?; |
663 | 22.9k | let ParsedItem(input, mut nanosecond) = |
664 | 22.9k | if let Some(ParsedItem(input, ())) = ascii_char::<b'.'>(input) { |
665 | 1.83k | let ParsedItem(mut input, mut value) = any_digit(input) |
666 | 1.83k | .ok_or(InvalidComponent("subsecond"))? |
667 | 1.80k | .map(|v| (v - b'0').extend::<u32>() * 100_000_000); |
668 | | |
669 | 1.80k | let mut multiplier = 10_000_000; |
670 | 7.47k | while let Some(ParsedItem(new_input, digit)) = any_digit(input) { |
671 | 5.67k | value += (digit - b'0').extend::<u32>() * multiplier; |
672 | 5.67k | input = new_input; |
673 | 5.67k | multiplier /= 10; |
674 | 5.67k | } |
675 | | |
676 | 1.80k | ParsedItem(input, value) |
677 | | } else { |
678 | 21.1k | ParsedItem(input, 0) |
679 | | }; |
680 | 22.8k | let ParsedItem(input, offset) = { |
681 | 22.9k | if let Some(ParsedItem(input, ())) = ascii_char_ignore_case::<b'Z'>(input) { |
682 | 2.11k | ParsedItem(input, UtcOffset::UTC) |
683 | | } else { |
684 | 20.7k | let ParsedItem(input, offset_sign) = |
685 | 20.7k | sign(input).ok_or(InvalidComponent("offset hour"))?; |
686 | 20.7k | let ParsedItem(input, offset_hour) = exactly_n_digits::<2, u8>(input) |
687 | 20.7k | .and_then(|parsed| parsed.filter(|&offset_hour| offset_hour <= 23)) |
688 | 20.7k | .ok_or(InvalidComponent("offset hour"))?; |
689 | 20.7k | let input = colon(input).ok_or(InvalidLiteral)?.into_inner(); |
690 | 20.6k | let ParsedItem(input, offset_minute) = |
691 | 20.7k | exactly_n_digits::<2, u8>(input).ok_or(InvalidComponent("offset minute"))?; |
692 | 20.6k | UtcOffset::from_hms( |
693 | 20.6k | if offset_sign == b'-' { |
694 | 3.38k | -offset_hour.cast_signed() |
695 | | } else { |
696 | 17.3k | offset_hour.cast_signed() |
697 | | }, |
698 | 20.6k | if offset_sign == b'-' { |
699 | 3.38k | -offset_minute.cast_signed() |
700 | | } else { |
701 | 17.3k | offset_minute.cast_signed() |
702 | | }, |
703 | | 0, |
704 | | ) |
705 | 20.6k | .map(|offset| ParsedItem(input, offset)) |
706 | 20.6k | .map_err(|mut err| { |
707 | | // Provide the user a more accurate error. |
708 | 6 | if err.name == "hours" { |
709 | 0 | err.name = "offset hour"; |
710 | 6 | } else if err.name == "minutes" { |
711 | 6 | err.name = "offset minute"; |
712 | 6 | } |
713 | 6 | err |
714 | 6 | }) |
715 | 20.6k | .map_err(TryFromParsed::ComponentRange)? |
716 | | } |
717 | | }; |
718 | | |
719 | 22.8k | if !input.is_empty() { |
720 | 19 | return Err(error::Parse::ParseFromDescription( |
721 | 19 | error::ParseFromDescription::UnexpectedTrailingCharacters, |
722 | 19 | )); |
723 | 22.7k | } |
724 | | |
725 | | // The RFC explicitly permits leap seconds. We don't currently support them, so treat it as |
726 | | // the preceding nanosecond. However, leap seconds can only occur as the last second of the |
727 | | // month UTC. |
728 | 22.7k | let leap_second_input = if second == 60 { |
729 | 11.5k | second = 59; |
730 | 11.5k | nanosecond = 999_999_999; |
731 | 11.5k | true |
732 | | } else { |
733 | 11.2k | false |
734 | | }; |
735 | | |
736 | 22.7k | let date = Month::from_number(month) |
737 | 22.7k | .and_then(|month| Date::from_calendar_date(year.cast_signed(), month, day)) |
738 | 22.7k | .map_err(TryFromParsed::ComponentRange)?; |
739 | 22.7k | let time = Time::from_hms_nano(hour, minute, second, nanosecond) |
740 | 22.7k | .map_err(TryFromParsed::ComponentRange)?; |
741 | 22.7k | let dt = OffsetDateTime::new_in_offset(date, time, offset); |
742 | | |
743 | 22.7k | if leap_second_input && !dt.is_valid_leap_second_stand_in() { |
744 | 104 | return Err(error::Parse::TryFromParsed(TryFromParsed::ComponentRange( |
745 | 104 | error::ComponentRange { |
746 | 104 | name: "second", |
747 | 104 | minimum: 0, |
748 | 104 | maximum: 59, |
749 | 104 | value: 60, |
750 | 104 | conditional_message: Some("because leap seconds are not supported"), |
751 | 104 | }, |
752 | 104 | ))); |
753 | 22.6k | } |
754 | | |
755 | 22.6k | Ok(dt) |
756 | 23.3k | } |
757 | | } |
758 | | |
759 | | impl<const CONFIG: EncodedConfig> sealed::Sealed for Iso8601<CONFIG> { |
760 | | #[inline] |
761 | 0 | fn parse_into<'a>( |
762 | 0 | &self, |
763 | 0 | mut input: &'a [u8], |
764 | 0 | parsed: &mut Parsed, |
765 | 0 | ) -> Result<&'a [u8], error::Parse> { |
766 | | use crate::parsing::combinator::rfc::iso8601::ExtendedKind; |
767 | | |
768 | 0 | let mut extended_kind = ExtendedKind::Unknown; |
769 | 0 | let mut date_is_present = false; |
770 | 0 | let mut time_is_present = false; |
771 | 0 | let mut offset_is_present = false; |
772 | 0 | let mut first_error = None; |
773 | | |
774 | 0 | parsed.leap_second_allowed = true; |
775 | | |
776 | 0 | match Self::parse_date(parsed, &mut extended_kind)(input) { |
777 | 0 | Ok(new_input) => { |
778 | 0 | input = new_input; |
779 | 0 | date_is_present = true; |
780 | 0 | } |
781 | 0 | Err(err) => { |
782 | 0 | first_error.get_or_insert(err); |
783 | 0 | } |
784 | | } |
785 | | |
786 | 0 | match Self::parse_time(parsed, &mut extended_kind, date_is_present)(input) { |
787 | 0 | Ok(new_input) => { |
788 | 0 | input = new_input; |
789 | 0 | time_is_present = true; |
790 | 0 | } |
791 | 0 | Err(err) => { |
792 | 0 | first_error.get_or_insert(err); |
793 | 0 | } |
794 | | } |
795 | | |
796 | | // If a date and offset are present, a time must be as well. |
797 | 0 | if !date_is_present || time_is_present { |
798 | 0 | match Self::parse_offset(parsed, &mut extended_kind)(input) { |
799 | 0 | Ok(new_input) => { |
800 | 0 | input = new_input; |
801 | 0 | offset_is_present = true; |
802 | 0 | } |
803 | 0 | Err(err) => { |
804 | 0 | first_error.get_or_insert(err); |
805 | 0 | } |
806 | | } |
807 | 0 | } |
808 | | |
809 | 0 | if !date_is_present && !time_is_present && !offset_is_present { |
810 | 0 | match first_error { |
811 | 0 | Some(err) => return Err(err), |
812 | 0 | None => bug!("an error should be present if no components were parsed"), |
813 | | } |
814 | 0 | } |
815 | | |
816 | 0 | Ok(input) |
817 | 0 | } |
818 | | } |