1 | // pest. The Elegant Parser |
2 | // Copyright (c) 2018 Dragoș Tiselice |
3 | // |
4 | // Licensed under the Apache License, Version 2.0 |
5 | // <LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0> or the MIT |
6 | // license <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your |
7 | // option. All files in the project carrying such notice may not be copied, |
8 | // modified, or distributed except according to those terms. |
9 | |
10 | //! Types for different kinds of parsing failures. |
11 | |
12 | use crate::parser_state::{ParseAttempts, ParsingToken, RulesCallStack}; |
13 | use alloc::borrow::Cow; |
14 | use alloc::borrow::ToOwned; |
15 | use alloc::boxed::Box; |
16 | use alloc::collections::{BTreeMap, BTreeSet}; |
17 | use alloc::format; |
18 | use alloc::string::String; |
19 | use alloc::string::ToString; |
20 | use alloc::vec; |
21 | use alloc::vec::Vec; |
22 | use core::cmp; |
23 | use core::fmt; |
24 | use core::mem; |
25 | |
26 | use crate::position::Position; |
27 | use crate::span::Span; |
28 | use crate::RuleType; |
29 | |
30 | /// Parse-related error type. |
31 | #[derive (Clone, Debug, Eq, Hash, PartialEq)] |
32 | #[cfg_attr (feature = "std" , derive(thiserror::Error))] |
33 | pub struct Error<R> { |
34 | /// Variant of the error |
35 | pub variant: ErrorVariant<R>, |
36 | /// Location within the input string |
37 | pub location: InputLocation, |
38 | /// Line/column within the input string |
39 | pub line_col: LineColLocation, |
40 | path: Option<String>, |
41 | line: String, |
42 | continued_line: Option<String>, |
43 | parse_attempts: Option<ParseAttempts<R>>, |
44 | } |
45 | |
46 | /// Different kinds of parsing errors. |
47 | #[derive (Clone, Debug, Eq, Hash, PartialEq)] |
48 | #[cfg_attr (feature = "std" , derive(thiserror::Error))] |
49 | pub enum ErrorVariant<R> { |
50 | /// Generated parsing error with expected and unexpected `Rule`s |
51 | ParsingError { |
52 | /// Positive attempts |
53 | positives: Vec<R>, |
54 | /// Negative attempts |
55 | negatives: Vec<R>, |
56 | }, |
57 | /// Custom error with a message |
58 | CustomError { |
59 | /// Short explanation |
60 | message: String, |
61 | }, |
62 | } |
63 | |
64 | /// Where an `Error` has occurred. |
65 | #[derive (Clone, Debug, Eq, Hash, PartialEq)] |
66 | pub enum InputLocation { |
67 | /// `Error` was created by `Error::new_from_pos` |
68 | Pos(usize), |
69 | /// `Error` was created by `Error::new_from_span` |
70 | Span((usize, usize)), |
71 | } |
72 | |
73 | /// Line/column where an `Error` has occurred. |
74 | #[derive (Clone, Debug, Eq, Hash, PartialEq)] |
75 | pub enum LineColLocation { |
76 | /// Line/column pair if `Error` was created by `Error::new_from_pos` |
77 | Pos((usize, usize)), |
78 | /// Line/column pairs if `Error` was created by `Error::new_from_span` |
79 | Span((usize, usize), (usize, usize)), |
80 | } |
81 | |
82 | impl From<Position<'_>> for LineColLocation { |
83 | fn from(value: Position<'_>) -> Self { |
84 | Self::Pos(value.line_col()) |
85 | } |
86 | } |
87 | |
88 | impl From<Span<'_>> for LineColLocation { |
89 | fn from(value: Span<'_>) -> Self { |
90 | let (start: Position<'_>, end: Position<'_>) = value.split(); |
91 | Self::Span(start.line_col(), end.line_col()) |
92 | } |
93 | } |
94 | |
95 | /// Function mapping rule to its helper message defined by user. |
96 | pub type RuleToMessageFn<R> = Box<dyn Fn(&R) -> Option<String>>; |
97 | /// Function mapping string element to bool denoting whether it's a whitespace defined by user. |
98 | pub type IsWhitespaceFn = Box<dyn Fn(String) -> bool>; |
99 | |
100 | impl ParsingToken { |
101 | pub fn is_whitespace(&self, is_whitespace: &IsWhitespaceFn) -> bool { |
102 | match self { |
103 | ParsingToken::Sensitive { token: &String } => is_whitespace(token.clone()), |
104 | ParsingToken::Insensitive { token: &String } => is_whitespace(token.clone()), |
105 | ParsingToken::Range { .. } => false, |
106 | ParsingToken::BuiltInRule => false, |
107 | } |
108 | } |
109 | } |
110 | |
111 | impl<R: RuleType> ParseAttempts<R> { |
112 | /// Helper formatting function to get message informing about tokens we've |
113 | /// (un)expected to see. |
114 | /// Used as a part of `parse_attempts_error`. |
115 | fn tokens_helper_messages( |
116 | &self, |
117 | is_whitespace_fn: &IsWhitespaceFn, |
118 | spacing: &str, |
119 | ) -> Vec<String> { |
120 | let mut helper_messages = Vec::new(); |
121 | let tokens_header_pairs = vec![ |
122 | (self.expected_tokens(), "expected" ), |
123 | (self.unexpected_tokens(), "unexpected" ), |
124 | ]; |
125 | |
126 | for (tokens, header) in &tokens_header_pairs { |
127 | if tokens.is_empty() { |
128 | continue; |
129 | } |
130 | |
131 | let mut helper_tokens_message = format!(" {spacing}note: {header} " ); |
132 | helper_tokens_message.push_str(if tokens.len() == 1 { |
133 | "token: " |
134 | } else { |
135 | "one of tokens: " |
136 | }); |
137 | |
138 | let expected_tokens_set: BTreeSet<String> = tokens |
139 | .iter() |
140 | .map(|token| { |
141 | if token.is_whitespace(is_whitespace_fn) { |
142 | String::from("WHITESPACE" ) |
143 | } else { |
144 | format!("` {}`" , token) |
145 | } |
146 | }) |
147 | .collect(); |
148 | |
149 | helper_tokens_message.push_str( |
150 | &expected_tokens_set |
151 | .iter() |
152 | .cloned() |
153 | .collect::<Vec<String>>() |
154 | .join(", " ), |
155 | ); |
156 | helper_messages.push(helper_tokens_message); |
157 | } |
158 | |
159 | helper_messages |
160 | } |
161 | } |
162 | |
163 | impl<R: RuleType> Error<R> { |
164 | /// Creates `Error` from `ErrorVariant` and `Position`. |
165 | /// |
166 | /// # Examples |
167 | /// |
168 | /// ``` |
169 | /// # use pest::error::{Error, ErrorVariant}; |
170 | /// # use pest::Position; |
171 | /// # #[allow (non_camel_case_types)] |
172 | /// # #[allow (dead_code)] |
173 | /// # #[derive(Clone, Copy, Debug, Eq, Hash, Ord, PartialEq, PartialOrd)] |
174 | /// # enum Rule { |
175 | /// # open_paren, |
176 | /// # closed_paren |
177 | /// # } |
178 | /// # let input = "" ; |
179 | /// # let pos = Position::from_start(input); |
180 | /// let error = Error::new_from_pos( |
181 | /// ErrorVariant::ParsingError { |
182 | /// positives: vec![Rule::open_paren], |
183 | /// negatives: vec![Rule::closed_paren], |
184 | /// }, |
185 | /// pos |
186 | /// ); |
187 | /// |
188 | /// println!("{}" , error); |
189 | /// ``` |
190 | pub fn new_from_pos(variant: ErrorVariant<R>, pos: Position<'_>) -> Error<R> { |
191 | let visualize_ws = pos.match_char(' \n' ) || pos.match_char(' \r' ); |
192 | let line_of = pos.line_of(); |
193 | let line = if visualize_ws { |
194 | visualize_whitespace(line_of) |
195 | } else { |
196 | line_of.replace(&[' \r' , ' \n' ][..], "" ) |
197 | }; |
198 | Error { |
199 | variant, |
200 | location: InputLocation::Pos(pos.pos()), |
201 | path: None, |
202 | line, |
203 | continued_line: None, |
204 | line_col: LineColLocation::Pos(pos.line_col()), |
205 | parse_attempts: None, |
206 | } |
207 | } |
208 | |
209 | /// Wrapper function to track `parse_attempts` as a result |
210 | /// of `state` function call in `parser_state.rs`. |
211 | pub(crate) fn new_from_pos_with_parsing_attempts( |
212 | variant: ErrorVariant<R>, |
213 | pos: Position<'_>, |
214 | parse_attempts: ParseAttempts<R>, |
215 | ) -> Error<R> { |
216 | let mut error = Self::new_from_pos(variant, pos); |
217 | error.parse_attempts = Some(parse_attempts); |
218 | error |
219 | } |
220 | |
221 | /// Creates `Error` from `ErrorVariant` and `Span`. |
222 | /// |
223 | /// # Examples |
224 | /// |
225 | /// ``` |
226 | /// # use pest::error::{Error, ErrorVariant}; |
227 | /// # use pest::{Position, Span}; |
228 | /// # #[allow (non_camel_case_types)] |
229 | /// # #[allow (dead_code)] |
230 | /// # #[derive(Clone, Copy, Debug, Eq, Hash, Ord, PartialEq, PartialOrd)] |
231 | /// # enum Rule { |
232 | /// # open_paren, |
233 | /// # closed_paren |
234 | /// # } |
235 | /// # let input = "" ; |
236 | /// # let start = Position::from_start(input); |
237 | /// # let end = start.clone(); |
238 | /// # let span = start.span(&end); |
239 | /// let error = Error::new_from_span( |
240 | /// ErrorVariant::ParsingError { |
241 | /// positives: vec![Rule::open_paren], |
242 | /// negatives: vec![Rule::closed_paren], |
243 | /// }, |
244 | /// span |
245 | /// ); |
246 | /// |
247 | /// println!("{}" , error); |
248 | /// ``` |
249 | pub fn new_from_span(variant: ErrorVariant<R>, span: Span<'_>) -> Error<R> { |
250 | let end = span.end_pos(); |
251 | let mut end_line_col = end.line_col(); |
252 | // end position is after a \n, so we want to point to the visual lf symbol |
253 | if end_line_col.1 == 1 { |
254 | let mut visual_end = end; |
255 | visual_end.skip_back(1); |
256 | let lc = visual_end.line_col(); |
257 | end_line_col = (lc.0, lc.1 + 1); |
258 | }; |
259 | |
260 | let mut line_iter = span.lines(); |
261 | let sl = line_iter.next().unwrap_or("" ); |
262 | let mut chars = span.as_str().chars(); |
263 | let visualize_ws = matches!(chars.next(), Some(' \n' ) | Some(' \r' )) |
264 | || matches!(chars.last(), Some(' \n' ) | Some(' \r' )); |
265 | let start_line = if visualize_ws { |
266 | visualize_whitespace(sl) |
267 | } else { |
268 | sl.to_owned().replace(&[' \r' , ' \n' ][..], "" ) |
269 | }; |
270 | let ll = line_iter.last(); |
271 | let continued_line = if visualize_ws { |
272 | ll.map(str::to_owned) |
273 | } else { |
274 | ll.map(visualize_whitespace) |
275 | }; |
276 | |
277 | Error { |
278 | variant, |
279 | location: InputLocation::Span((span.start(), end.pos())), |
280 | path: None, |
281 | line: start_line, |
282 | continued_line, |
283 | line_col: LineColLocation::Span(span.start_pos().line_col(), end_line_col), |
284 | parse_attempts: None, |
285 | } |
286 | } |
287 | |
288 | /// Returns `Error` variant with `path` which is shown when formatted with `Display`. |
289 | /// |
290 | /// # Examples |
291 | /// |
292 | /// ``` |
293 | /// # use pest::error::{Error, ErrorVariant}; |
294 | /// # use pest::Position; |
295 | /// # #[allow (non_camel_case_types)] |
296 | /// # #[allow (dead_code)] |
297 | /// # #[derive(Clone, Copy, Debug, Eq, Hash, Ord, PartialEq, PartialOrd)] |
298 | /// # enum Rule { |
299 | /// # open_paren, |
300 | /// # closed_paren |
301 | /// # } |
302 | /// # let input = "" ; |
303 | /// # let pos = Position::from_start(input); |
304 | /// Error::new_from_pos( |
305 | /// ErrorVariant::ParsingError { |
306 | /// positives: vec![Rule::open_paren], |
307 | /// negatives: vec![Rule::closed_paren], |
308 | /// }, |
309 | /// pos |
310 | /// ).with_path("file.rs" ); |
311 | /// ``` |
312 | pub fn with_path(mut self, path: &str) -> Error<R> { |
313 | self.path = Some(path.to_owned()); |
314 | |
315 | self |
316 | } |
317 | |
318 | /// Returns the path set using [`Error::with_path()`]. |
319 | /// |
320 | /// # Examples |
321 | /// |
322 | /// ``` |
323 | /// # use pest::error::{Error, ErrorVariant}; |
324 | /// # use pest::Position; |
325 | /// # #[allow (non_camel_case_types)] |
326 | /// # #[allow (dead_code)] |
327 | /// # #[derive(Clone, Copy, Debug, Eq, Hash, Ord, PartialEq, PartialOrd)] |
328 | /// # enum Rule { |
329 | /// # open_paren, |
330 | /// # closed_paren |
331 | /// # } |
332 | /// # let input = "" ; |
333 | /// # let pos = Position::from_start(input); |
334 | /// # let error = Error::new_from_pos( |
335 | /// # ErrorVariant::ParsingError { |
336 | /// # positives: vec![Rule::open_paren], |
337 | /// # negatives: vec![Rule::closed_paren], |
338 | /// # }, |
339 | /// # pos); |
340 | /// let error = error.with_path("file.rs" ); |
341 | /// assert_eq!(Some("file.rs" ), error.path()); |
342 | /// ``` |
343 | pub fn path(&self) -> Option<&str> { |
344 | self.path.as_deref() |
345 | } |
346 | |
347 | /// Returns the line that the error is on. |
348 | pub fn line(&self) -> &str { |
349 | self.line.as_str() |
350 | } |
351 | |
352 | /// Renames all `Rule`s if this is a [`ParsingError`]. It does nothing when called on a |
353 | /// [`CustomError`]. |
354 | /// |
355 | /// Useful in order to rename verbose rules or have detailed per-`Rule` formatting. |
356 | /// |
357 | /// [`ParsingError`]: enum.ErrorVariant.html#variant.ParsingError |
358 | /// [`CustomError`]: enum.ErrorVariant.html#variant.CustomError |
359 | /// |
360 | /// # Examples |
361 | /// |
362 | /// ``` |
363 | /// # use pest::error::{Error, ErrorVariant}; |
364 | /// # use pest::Position; |
365 | /// # #[allow (non_camel_case_types)] |
366 | /// # #[allow (dead_code)] |
367 | /// # #[derive(Clone, Copy, Debug, Eq, Hash, Ord, PartialEq, PartialOrd)] |
368 | /// # enum Rule { |
369 | /// # open_paren, |
370 | /// # closed_paren |
371 | /// # } |
372 | /// # let input = "" ; |
373 | /// # let pos = Position::from_start(input); |
374 | /// Error::new_from_pos( |
375 | /// ErrorVariant::ParsingError { |
376 | /// positives: vec![Rule::open_paren], |
377 | /// negatives: vec![Rule::closed_paren], |
378 | /// }, |
379 | /// pos |
380 | /// ).renamed_rules(|rule| { |
381 | /// match *rule { |
382 | /// Rule::open_paren => "(" .to_owned(), |
383 | /// Rule::closed_paren => "closed paren" .to_owned() |
384 | /// } |
385 | /// }); |
386 | /// ``` |
387 | pub fn renamed_rules<F>(mut self, f: F) -> Error<R> |
388 | where |
389 | F: FnMut(&R) -> String, |
390 | { |
391 | let variant = match self.variant { |
392 | ErrorVariant::ParsingError { |
393 | positives, |
394 | negatives, |
395 | } => { |
396 | let message = Error::parsing_error_message(&positives, &negatives, f); |
397 | ErrorVariant::CustomError { message } |
398 | } |
399 | variant => variant, |
400 | }; |
401 | |
402 | self.variant = variant; |
403 | |
404 | self |
405 | } |
406 | |
407 | /// Get detailed information about errored rules sequence. |
408 | /// Returns `Some(results)` only for `ParsingError`. |
409 | pub fn parse_attempts(&self) -> Option<ParseAttempts<R>> { |
410 | self.parse_attempts.clone() |
411 | } |
412 | |
413 | /// Get error message based on parsing attempts. |
414 | /// Returns `None` in case self `parse_attempts` is `None`. |
415 | pub fn parse_attempts_error( |
416 | &self, |
417 | input: &str, |
418 | rule_to_message: &RuleToMessageFn<R>, |
419 | is_whitespace: &IsWhitespaceFn, |
420 | ) -> Option<Error<R>> { |
421 | let attempts = if let Some(ref parse_attempts) = self.parse_attempts { |
422 | parse_attempts.clone() |
423 | } else { |
424 | return None; |
425 | }; |
426 | |
427 | let spacing = self.spacing() + " " ; |
428 | let error_position = attempts.max_position; |
429 | let message = { |
430 | let mut help_lines: Vec<String> = Vec::new(); |
431 | help_lines.push(String::from("error: parsing error occurred." )); |
432 | |
433 | // Note: at least one of `(un)expected_tokens` must not be empty. |
434 | for tokens_helper_message in attempts.tokens_helper_messages(is_whitespace, &spacing) { |
435 | help_lines.push(tokens_helper_message); |
436 | } |
437 | |
438 | let call_stacks = attempts.call_stacks(); |
439 | // Group call stacks by their parents so that we can print common header and |
440 | // several sub helper messages. |
441 | let mut call_stacks_parents_groups: BTreeMap<Option<R>, Vec<RulesCallStack<R>>> = |
442 | BTreeMap::new(); |
443 | for call_stack in call_stacks { |
444 | call_stacks_parents_groups |
445 | .entry(call_stack.parent) |
446 | .or_default() |
447 | .push(call_stack); |
448 | } |
449 | |
450 | for (group_parent, group) in call_stacks_parents_groups { |
451 | if let Some(parent_rule) = group_parent { |
452 | let mut contains_meaningful_info = false; |
453 | help_lines.push(format!( |
454 | " {spacing}help: {}" , |
455 | if let Some(message) = rule_to_message(&parent_rule) { |
456 | contains_meaningful_info = true; |
457 | message |
458 | } else { |
459 | String::from("[Unknown parent rule]" ) |
460 | } |
461 | )); |
462 | for call_stack in group { |
463 | if let Some(r) = call_stack.deepest.get_rule() { |
464 | if let Some(message) = rule_to_message(r) { |
465 | contains_meaningful_info = true; |
466 | help_lines.push(format!(" {spacing} - {message}" )); |
467 | } |
468 | } |
469 | } |
470 | if !contains_meaningful_info { |
471 | // Have to remove useless line for unknown parent rule. |
472 | help_lines.pop(); |
473 | } |
474 | } else { |
475 | for call_stack in group { |
476 | // Note that `deepest` rule may be `None`. E.g. in case it corresponds |
477 | // to WHITESPACE expected token which has no parent rule (on the top level |
478 | // parsing). |
479 | if let Some(r) = call_stack.deepest.get_rule() { |
480 | let helper_message = rule_to_message(r); |
481 | if let Some(helper_message) = helper_message { |
482 | help_lines.push(format!(" {spacing}help: {helper_message}" )); |
483 | } |
484 | } |
485 | } |
486 | } |
487 | } |
488 | |
489 | help_lines.join(" \n" ) |
490 | }; |
491 | let error = Error::new_from_pos( |
492 | ErrorVariant::CustomError { message }, |
493 | Position::new_internal(input, error_position), |
494 | ); |
495 | Some(error) |
496 | } |
497 | |
498 | fn start(&self) -> (usize, usize) { |
499 | match self.line_col { |
500 | LineColLocation::Pos(line_col) => line_col, |
501 | LineColLocation::Span(start_line_col, _) => start_line_col, |
502 | } |
503 | } |
504 | |
505 | fn spacing(&self) -> String { |
506 | let line = match self.line_col { |
507 | LineColLocation::Pos((line, _)) => line, |
508 | LineColLocation::Span((start_line, _), (end_line, _)) => cmp::max(start_line, end_line), |
509 | }; |
510 | |
511 | let line_str_len = format!(" {}" , line).len(); |
512 | |
513 | let mut spacing = String::new(); |
514 | for _ in 0..line_str_len { |
515 | spacing.push(' ' ); |
516 | } |
517 | |
518 | spacing |
519 | } |
520 | |
521 | fn underline(&self) -> String { |
522 | let mut underline = String::new(); |
523 | |
524 | let mut start = self.start().1; |
525 | let end = match self.line_col { |
526 | LineColLocation::Span(_, (_, mut end)) => { |
527 | let inverted_cols = start > end; |
528 | if inverted_cols { |
529 | mem::swap(&mut start, &mut end); |
530 | start -= 1; |
531 | end += 1; |
532 | } |
533 | |
534 | Some(end) |
535 | } |
536 | _ => None, |
537 | }; |
538 | let offset = start - 1; |
539 | let line_chars = self.line.chars(); |
540 | |
541 | for c in line_chars.take(offset) { |
542 | match c { |
543 | ' \t' => underline.push(' \t' ), |
544 | _ => underline.push(' ' ), |
545 | } |
546 | } |
547 | |
548 | if let Some(end) = end { |
549 | underline.push('^' ); |
550 | if end - start > 1 { |
551 | for _ in 2..(end - start) { |
552 | underline.push('-' ); |
553 | } |
554 | underline.push('^' ); |
555 | } |
556 | } else { |
557 | underline.push_str("^---" ) |
558 | } |
559 | |
560 | underline |
561 | } |
562 | |
563 | fn message(&self) -> String { |
564 | self.variant.message().to_string() |
565 | } |
566 | |
567 | fn parsing_error_message<F>(positives: &[R], negatives: &[R], mut f: F) -> String |
568 | where |
569 | F: FnMut(&R) -> String, |
570 | { |
571 | match (negatives.is_empty(), positives.is_empty()) { |
572 | (false, false) => format!( |
573 | "unexpected {}; expected {}" , |
574 | Error::enumerate(negatives, &mut f), |
575 | Error::enumerate(positives, &mut f) |
576 | ), |
577 | (false, true) => format!("unexpected {}" , Error::enumerate(negatives, &mut f)), |
578 | (true, false) => format!("expected {}" , Error::enumerate(positives, &mut f)), |
579 | (true, true) => "unknown parsing error" .to_owned(), |
580 | } |
581 | } |
582 | |
583 | fn enumerate<F>(rules: &[R], f: &mut F) -> String |
584 | where |
585 | F: FnMut(&R) -> String, |
586 | { |
587 | match rules.len() { |
588 | 1 => f(&rules[0]), |
589 | 2 => format!(" {} or {}" , f(&rules[0]), f(&rules[1])), |
590 | l => { |
591 | let non_separated = f(&rules[l - 1]); |
592 | let separated = rules |
593 | .iter() |
594 | .take(l - 1) |
595 | .map(f) |
596 | .collect::<Vec<_>>() |
597 | .join(", " ); |
598 | format!(" {}, or {}" , separated, non_separated) |
599 | } |
600 | } |
601 | } |
602 | |
603 | pub(crate) fn format(&self) -> String { |
604 | let spacing = self.spacing(); |
605 | let path = self |
606 | .path |
607 | .as_ref() |
608 | .map(|path| format!(" {}:" , path)) |
609 | .unwrap_or_default(); |
610 | |
611 | let pair = (self.line_col.clone(), &self.continued_line); |
612 | if let (LineColLocation::Span(_, end), Some(ref continued_line)) = pair { |
613 | let has_line_gap = end.0 - self.start().0 > 1; |
614 | if has_line_gap { |
615 | format!( |
616 | " {s }--> {p}{ls}: {c}\n\ |
617 | {s } | \n\ |
618 | {ls:w$} | {line}\n\ |
619 | {s } | ... \n\ |
620 | {le:w$} | {continued_line}\n\ |
621 | {s } | {underline}\n\ |
622 | {s } | \n\ |
623 | {s } = {message}" , |
624 | s = spacing, |
625 | w = spacing.len(), |
626 | p = path, |
627 | ls = self.start().0, |
628 | le = end.0, |
629 | c = self.start().1, |
630 | line = self.line, |
631 | continued_line = continued_line, |
632 | underline = self.underline(), |
633 | message = self.message() |
634 | ) |
635 | } else { |
636 | format!( |
637 | " {s }--> {p}{ls}: {c}\n\ |
638 | {s } | \n\ |
639 | {ls:w$} | {line}\n\ |
640 | {le:w$} | {continued_line}\n\ |
641 | {s } | {underline}\n\ |
642 | {s } | \n\ |
643 | {s } = {message}" , |
644 | s = spacing, |
645 | w = spacing.len(), |
646 | p = path, |
647 | ls = self.start().0, |
648 | le = end.0, |
649 | c = self.start().1, |
650 | line = self.line, |
651 | continued_line = continued_line, |
652 | underline = self.underline(), |
653 | message = self.message() |
654 | ) |
655 | } |
656 | } else { |
657 | format!( |
658 | " {s}--> {p}{l}: {c}\n\ |
659 | {s} | \n\ |
660 | {l} | {line}\n\ |
661 | {s} | {underline}\n\ |
662 | {s} | \n\ |
663 | {s} = {message}" , |
664 | s = spacing, |
665 | p = path, |
666 | l = self.start().0, |
667 | c = self.start().1, |
668 | line = self.line, |
669 | underline = self.underline(), |
670 | message = self.message() |
671 | ) |
672 | } |
673 | } |
674 | |
675 | #[cfg (feature = "miette-error" )] |
676 | /// Turns an error into a [miette](crates.io/miette) Diagnostic. |
677 | pub fn into_miette(self) -> impl ::miette::Diagnostic { |
678 | miette_adapter::MietteAdapter(self) |
679 | } |
680 | } |
681 | |
682 | impl<R: RuleType> ErrorVariant<R> { |
683 | /// |
684 | /// Returns the error message for [`ErrorVariant`] |
685 | /// |
686 | /// If [`ErrorVariant`] is [`CustomError`], it returns a |
687 | /// [`Cow::Borrowed`] reference to [`message`]. If [`ErrorVariant`] is [`ParsingError`], a |
688 | /// [`Cow::Owned`] containing "expected [ErrorVariant::ParsingError::positives] [ErrorVariant::ParsingError::negatives]" is returned. |
689 | /// |
690 | /// [`ErrorVariant`]: enum.ErrorVariant.html |
691 | /// [`CustomError`]: enum.ErrorVariant.html#variant.CustomError |
692 | /// [`ParsingError`]: enum.ErrorVariant.html#variant.ParsingError |
693 | /// [`Cow::Owned`]: https://doc.rust-lang.org/std/borrow/enum.Cow.html#variant.Owned |
694 | /// [`Cow::Borrowed`]: https://doc.rust-lang.org/std/borrow/enum.Cow.html#variant.Borrowed |
695 | /// [`message`]: enum.ErrorVariant.html#variant.CustomError.field.message |
696 | /// # Examples |
697 | /// |
698 | /// ``` |
699 | /// # use pest::error::ErrorVariant; |
700 | /// let variant = ErrorVariant::<()>::CustomError { |
701 | /// message: String::from("unexpected error" ) |
702 | /// }; |
703 | /// |
704 | /// println!("{}" , variant.message()); |
705 | pub fn message(&self) -> Cow<'_, str> { |
706 | match self { |
707 | ErrorVariant::ParsingError { |
708 | ref positives, |
709 | ref negatives, |
710 | } => Cow::Owned(Error::parsing_error_message(positives, negatives, |r| { |
711 | format!(" {:?}" , r) |
712 | })), |
713 | ErrorVariant::CustomError { ref message } => Cow::Borrowed(message), |
714 | } |
715 | } |
716 | } |
717 | |
718 | impl<R: RuleType> fmt::Display for Error<R> { |
719 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
720 | write!(f, " {}" , self.format()) |
721 | } |
722 | } |
723 | |
724 | impl<R: RuleType> fmt::Display for ErrorVariant<R> { |
725 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
726 | match self { |
727 | ErrorVariant::ParsingError { .. } => write!(f, "parsing error: {}" , self.message()), |
728 | ErrorVariant::CustomError { .. } => write!(f, " {}" , self.message()), |
729 | } |
730 | } |
731 | } |
732 | |
733 | fn visualize_whitespace(input: &str) -> String { |
734 | input.to_owned().replace(' \r' , "␍" ).replace(from:' \n' , to:"␊" ) |
735 | } |
736 | |
737 | #[cfg (feature = "miette-error" )] |
738 | mod miette_adapter { |
739 | use alloc::string::ToString; |
740 | use std::boxed::Box; |
741 | |
742 | use crate::error::LineColLocation; |
743 | |
744 | use super::{Error, RuleType}; |
745 | |
746 | use miette::{Diagnostic, LabeledSpan, SourceCode}; |
747 | |
748 | #[derive (thiserror::Error, Debug)] |
749 | #[error("Failure to parse at {:?}" , self.0.line_col)] |
750 | pub(crate) struct MietteAdapter<R: RuleType>(pub(crate) Error<R>); |
751 | |
752 | impl<R: RuleType> Diagnostic for MietteAdapter<R> { |
753 | fn source_code(&self) -> Option<&dyn SourceCode> { |
754 | Some(&self.0.line) |
755 | } |
756 | |
757 | fn labels(&self) -> Option<Box<dyn Iterator<Item = LabeledSpan>>> { |
758 | let message = self.0.variant.message().to_string(); |
759 | |
760 | let (offset, length) = match self.0.line_col { |
761 | LineColLocation::Pos((_, c)) => (c - 1, 1), |
762 | LineColLocation::Span((_, start_c), (_, end_c)) => { |
763 | (start_c - 1, end_c - start_c + 1) |
764 | } |
765 | }; |
766 | |
767 | let span = LabeledSpan::new(Some(message), offset, length); |
768 | |
769 | Some(Box::new(std::iter::once(span))) |
770 | } |
771 | |
772 | fn help<'a>(&'a self) -> Option<Box<dyn core::fmt::Display + 'a>> { |
773 | Some(Box::new(self.0.message())) |
774 | } |
775 | } |
776 | } |
777 | |
778 | #[cfg (test)] |
779 | mod tests { |
780 | use super::*; |
781 | use alloc::vec; |
782 | |
783 | #[test ] |
784 | fn display_parsing_error_mixed() { |
785 | let input = "ab \ncd \nef" ; |
786 | let pos = Position::new(input, 4).unwrap(); |
787 | let error: Error<u32> = Error::new_from_pos( |
788 | ErrorVariant::ParsingError { |
789 | positives: vec![1, 2, 3], |
790 | negatives: vec![4, 5, 6], |
791 | }, |
792 | pos, |
793 | ); |
794 | |
795 | assert_eq!( |
796 | format!("{}" , error), |
797 | [ |
798 | " --> 2:2" , |
799 | " |" , |
800 | "2 | cd" , |
801 | " | ^---" , |
802 | " |" , |
803 | " = unexpected 4, 5, or 6; expected 1, 2, or 3" |
804 | ] |
805 | .join(" \n" ) |
806 | ); |
807 | } |
808 | |
809 | #[test ] |
810 | fn display_parsing_error_positives() { |
811 | let input = "ab \ncd \nef" ; |
812 | let pos = Position::new(input, 4).unwrap(); |
813 | let error: Error<u32> = Error::new_from_pos( |
814 | ErrorVariant::ParsingError { |
815 | positives: vec![1, 2], |
816 | negatives: vec![], |
817 | }, |
818 | pos, |
819 | ); |
820 | |
821 | assert_eq!( |
822 | format!("{}" , error), |
823 | [ |
824 | " --> 2:2" , |
825 | " |" , |
826 | "2 | cd" , |
827 | " | ^---" , |
828 | " |" , |
829 | " = expected 1 or 2" |
830 | ] |
831 | .join(" \n" ) |
832 | ); |
833 | } |
834 | |
835 | #[test ] |
836 | fn display_parsing_error_negatives() { |
837 | let input = "ab \ncd \nef" ; |
838 | let pos = Position::new(input, 4).unwrap(); |
839 | let error: Error<u32> = Error::new_from_pos( |
840 | ErrorVariant::ParsingError { |
841 | positives: vec![], |
842 | negatives: vec![4, 5, 6], |
843 | }, |
844 | pos, |
845 | ); |
846 | |
847 | assert_eq!( |
848 | format!("{}" , error), |
849 | [ |
850 | " --> 2:2" , |
851 | " |" , |
852 | "2 | cd" , |
853 | " | ^---" , |
854 | " |" , |
855 | " = unexpected 4, 5, or 6" |
856 | ] |
857 | .join(" \n" ) |
858 | ); |
859 | } |
860 | |
861 | #[test ] |
862 | fn display_parsing_error_unknown() { |
863 | let input = "ab \ncd \nef" ; |
864 | let pos = Position::new(input, 4).unwrap(); |
865 | let error: Error<u32> = Error::new_from_pos( |
866 | ErrorVariant::ParsingError { |
867 | positives: vec![], |
868 | negatives: vec![], |
869 | }, |
870 | pos, |
871 | ); |
872 | |
873 | assert_eq!( |
874 | format!("{}" , error), |
875 | [ |
876 | " --> 2:2" , |
877 | " |" , |
878 | "2 | cd" , |
879 | " | ^---" , |
880 | " |" , |
881 | " = unknown parsing error" |
882 | ] |
883 | .join(" \n" ) |
884 | ); |
885 | } |
886 | |
887 | #[test ] |
888 | fn display_custom_pos() { |
889 | let input = "ab \ncd \nef" ; |
890 | let pos = Position::new(input, 4).unwrap(); |
891 | let error: Error<u32> = Error::new_from_pos( |
892 | ErrorVariant::CustomError { |
893 | message: "error: big one" .to_owned(), |
894 | }, |
895 | pos, |
896 | ); |
897 | |
898 | assert_eq!( |
899 | format!("{}" , error), |
900 | [ |
901 | " --> 2:2" , |
902 | " |" , |
903 | "2 | cd" , |
904 | " | ^---" , |
905 | " |" , |
906 | " = error: big one" |
907 | ] |
908 | .join(" \n" ) |
909 | ); |
910 | } |
911 | |
912 | #[test ] |
913 | fn display_custom_span_two_lines() { |
914 | let input = "ab \ncd \nefgh" ; |
915 | let start = Position::new(input, 4).unwrap(); |
916 | let end = Position::new(input, 9).unwrap(); |
917 | let error: Error<u32> = Error::new_from_span( |
918 | ErrorVariant::CustomError { |
919 | message: "error: big one" .to_owned(), |
920 | }, |
921 | start.span(&end), |
922 | ); |
923 | |
924 | assert_eq!( |
925 | format!("{}" , error), |
926 | [ |
927 | " --> 2:2" , |
928 | " |" , |
929 | "2 | cd" , |
930 | "3 | efgh" , |
931 | " | ^^" , |
932 | " |" , |
933 | " = error: big one" |
934 | ] |
935 | .join(" \n" ) |
936 | ); |
937 | } |
938 | |
939 | #[test ] |
940 | fn display_custom_span_three_lines() { |
941 | let input = "ab \ncd \nefgh" ; |
942 | let start = Position::new(input, 1).unwrap(); |
943 | let end = Position::new(input, 9).unwrap(); |
944 | let error: Error<u32> = Error::new_from_span( |
945 | ErrorVariant::CustomError { |
946 | message: "error: big one" .to_owned(), |
947 | }, |
948 | start.span(&end), |
949 | ); |
950 | |
951 | assert_eq!( |
952 | format!("{}" , error), |
953 | [ |
954 | " --> 1:2" , |
955 | " |" , |
956 | "1 | ab" , |
957 | " | ..." , |
958 | "3 | efgh" , |
959 | " | ^^" , |
960 | " |" , |
961 | " = error: big one" |
962 | ] |
963 | .join(" \n" ) |
964 | ); |
965 | } |
966 | |
967 | #[test ] |
968 | fn display_custom_span_two_lines_inverted_cols() { |
969 | let input = "abcdef \ngh" ; |
970 | let start = Position::new(input, 5).unwrap(); |
971 | let end = Position::new(input, 8).unwrap(); |
972 | let error: Error<u32> = Error::new_from_span( |
973 | ErrorVariant::CustomError { |
974 | message: "error: big one" .to_owned(), |
975 | }, |
976 | start.span(&end), |
977 | ); |
978 | |
979 | assert_eq!( |
980 | format!("{}" , error), |
981 | [ |
982 | " --> 1:6" , |
983 | " |" , |
984 | "1 | abcdef" , |
985 | "2 | gh" , |
986 | " | ^----^" , |
987 | " |" , |
988 | " = error: big one" |
989 | ] |
990 | .join(" \n" ) |
991 | ); |
992 | } |
993 | |
994 | #[test ] |
995 | fn display_custom_span_end_after_newline() { |
996 | let input = "abcdef \n" ; |
997 | let start = Position::new(input, 0).unwrap(); |
998 | let end = Position::new(input, 7).unwrap(); |
999 | assert!(start.at_start()); |
1000 | assert!(end.at_end()); |
1001 | |
1002 | let error: Error<u32> = Error::new_from_span( |
1003 | ErrorVariant::CustomError { |
1004 | message: "error: big one" .to_owned(), |
1005 | }, |
1006 | start.span(&end), |
1007 | ); |
1008 | |
1009 | assert_eq!( |
1010 | format!("{}" , error), |
1011 | [ |
1012 | " --> 1:1" , |
1013 | " |" , |
1014 | "1 | abcdef␊" , |
1015 | " | ^-----^" , |
1016 | " |" , |
1017 | " = error: big one" |
1018 | ] |
1019 | .join(" \n" ) |
1020 | ); |
1021 | } |
1022 | |
1023 | #[test ] |
1024 | fn display_custom_span_empty() { |
1025 | let input = "" ; |
1026 | let start = Position::new(input, 0).unwrap(); |
1027 | let end = Position::new(input, 0).unwrap(); |
1028 | assert!(start.at_start()); |
1029 | assert!(end.at_end()); |
1030 | |
1031 | let error: Error<u32> = Error::new_from_span( |
1032 | ErrorVariant::CustomError { |
1033 | message: "error: empty" .to_owned(), |
1034 | }, |
1035 | start.span(&end), |
1036 | ); |
1037 | |
1038 | assert_eq!( |
1039 | format!("{}" , error), |
1040 | [ |
1041 | " --> 1:1" , |
1042 | " |" , |
1043 | "1 | " , |
1044 | " | ^" , |
1045 | " |" , |
1046 | " = error: empty" |
1047 | ] |
1048 | .join(" \n" ) |
1049 | ); |
1050 | } |
1051 | |
1052 | #[test ] |
1053 | fn mapped_parsing_error() { |
1054 | let input = "ab \ncd \nef" ; |
1055 | let pos = Position::new(input, 4).unwrap(); |
1056 | let error: Error<u32> = Error::new_from_pos( |
1057 | ErrorVariant::ParsingError { |
1058 | positives: vec![1, 2, 3], |
1059 | negatives: vec![4, 5, 6], |
1060 | }, |
1061 | pos, |
1062 | ) |
1063 | .renamed_rules(|n| format!("{}" , n + 1)); |
1064 | |
1065 | assert_eq!( |
1066 | format!("{}" , error), |
1067 | [ |
1068 | " --> 2:2" , |
1069 | " |" , |
1070 | "2 | cd" , |
1071 | " | ^---" , |
1072 | " |" , |
1073 | " = unexpected 5, 6, or 7; expected 2, 3, or 4" |
1074 | ] |
1075 | .join(" \n" ) |
1076 | ); |
1077 | } |
1078 | |
1079 | #[test ] |
1080 | fn error_with_path() { |
1081 | let input = "ab \ncd \nef" ; |
1082 | let pos = Position::new(input, 4).unwrap(); |
1083 | let error: Error<u32> = Error::new_from_pos( |
1084 | ErrorVariant::ParsingError { |
1085 | positives: vec![1, 2, 3], |
1086 | negatives: vec![4, 5, 6], |
1087 | }, |
1088 | pos, |
1089 | ) |
1090 | .with_path("file.rs" ); |
1091 | |
1092 | assert_eq!( |
1093 | format!("{}" , error), |
1094 | [ |
1095 | " --> file.rs:2:2" , |
1096 | " |" , |
1097 | "2 | cd" , |
1098 | " | ^---" , |
1099 | " |" , |
1100 | " = unexpected 4, 5, or 6; expected 1, 2, or 3" |
1101 | ] |
1102 | .join(" \n" ) |
1103 | ); |
1104 | } |
1105 | |
1106 | #[test ] |
1107 | fn underline_with_tabs() { |
1108 | let input = "a \txbc" ; |
1109 | let pos = Position::new(input, 2).unwrap(); |
1110 | let error: Error<u32> = Error::new_from_pos( |
1111 | ErrorVariant::ParsingError { |
1112 | positives: vec![1, 2, 3], |
1113 | negatives: vec![4, 5, 6], |
1114 | }, |
1115 | pos, |
1116 | ) |
1117 | .with_path("file.rs" ); |
1118 | |
1119 | assert_eq!( |
1120 | format!("{}" , error), |
1121 | [ |
1122 | " --> file.rs:1:3" , |
1123 | " |" , |
1124 | "1 | a xbc" , |
1125 | " | ^---" , |
1126 | " |" , |
1127 | " = unexpected 4, 5, or 6; expected 1, 2, or 3" |
1128 | ] |
1129 | .join(" \n" ) |
1130 | ); |
1131 | } |
1132 | |
1133 | #[test ] |
1134 | fn pos_to_lcl_conversion() { |
1135 | let input = "input" ; |
1136 | |
1137 | let pos = Position::new(input, 2).unwrap(); |
1138 | |
1139 | assert_eq!(LineColLocation::Pos(pos.line_col()), pos.into()); |
1140 | } |
1141 | |
1142 | #[test ] |
1143 | fn span_to_lcl_conversion() { |
1144 | let input = "input" ; |
1145 | |
1146 | let span = Span::new(input, 2, 4).unwrap(); |
1147 | let (start, end) = span.split(); |
1148 | |
1149 | assert_eq!( |
1150 | LineColLocation::Span(start.line_col(), end.line_col()), |
1151 | span.into() |
1152 | ); |
1153 | } |
1154 | |
1155 | #[cfg (feature = "miette-error" )] |
1156 | #[test ] |
1157 | fn miette_error() { |
1158 | let input = "abc \ndef" ; |
1159 | let pos = Position::new(input, 4).unwrap(); |
1160 | let error: Error<u32> = Error::new_from_pos( |
1161 | ErrorVariant::ParsingError { |
1162 | positives: vec![1, 2, 3], |
1163 | negatives: vec![4, 5, 6], |
1164 | }, |
1165 | pos, |
1166 | ); |
1167 | |
1168 | let miette_error = miette::Error::new(error.into_miette()); |
1169 | |
1170 | assert_eq!( |
1171 | format!("{:?}" , miette_error), |
1172 | [ |
1173 | "" , |
1174 | " \u{1b}[31m× \u{1b}[0m Failure to parse at Pos((2, 1))" , |
1175 | " ╭────" , |
1176 | " \u{1b}[2m1 \u{1b}[0m │ def" , |
1177 | " · \u{1b}[35;1m┬ \u{1b}[0m" , |
1178 | " · \u{1b}[35;1m╰── \u{1b}[35;1munexpected 4, 5, or 6; expected 1, 2, or 3 \u{1b}[0m \u{1b}[0m" , |
1179 | " ╰────" , |
1180 | " \u{1b}[36m help: \u{1b}[0munexpected 4, 5, or 6; expected 1, 2, or 3 \n" |
1181 | ] |
1182 | .join(" \n" ) |
1183 | ); |
1184 | } |
1185 | } |
1186 | |