1use alacritty_terminal::{
2 Term,
3 event::EventListener,
4 grid::Dimensions,
5 index::{Boundary, Column, Direction as AlacDirection, Line, Point as AlacPoint},
6 term::search::{Match, RegexIter, RegexSearch},
7};
8use regex::Regex;
9use std::{ops::Index, sync::LazyLock};
10
11const URL_REGEX: &str = r#"(ipfs:|ipns:|magnet:|mailto:|gemini://|gopher://|https://|http://|news:|file://|git://|ssh:|ftp://)[^\u{0000}-\u{001F}\u{007F}-\u{009F}<>"\s{-}\^⟨⟩`]+"#;
12// Optional suffix matches MSBuild diagnostic suffixes for path parsing in PathLikeWithPosition
13// https://learn.microsoft.com/en-us/visualstudio/msbuild/msbuild-diagnostic-format-for-tasks
14const WORD_REGEX: &str =
15 r#"[\$\+\w.\[\]:/\\@\-~()]+(?:\((?:\d+|\d+,\d+)\))|[\$\+\w.\[\]:/\\@\-~()]+"#;
16
17const PYTHON_FILE_LINE_REGEX: &str = r#"File "(?P<file>[^"]+)", line (?P<line>\d+)"#;
18
19static PYTHON_FILE_LINE_MATCHER: LazyLock<Regex> =
20 LazyLock::new(|| Regex::new(PYTHON_FILE_LINE_REGEX).unwrap());
21
22fn python_extract_path_and_line(input: &str) -> Option<(&str, u32)> {
23 if let Some(captures) = PYTHON_FILE_LINE_MATCHER.captures(input) {
24 let path_part = captures.name("file")?.as_str();
25
26 let line_number: u32 = captures.name("line")?.as_str().parse().ok()?;
27 return Some((path_part, line_number));
28 }
29 None
30}
31
32pub(super) struct RegexSearches {
33 url_regex: RegexSearch,
34 word_regex: RegexSearch,
35 python_file_line_regex: RegexSearch,
36}
37
38impl RegexSearches {
39 pub(super) fn new() -> Self {
40 Self {
41 url_regex: RegexSearch::new(URL_REGEX).unwrap(),
42 word_regex: RegexSearch::new(WORD_REGEX).unwrap(),
43 python_file_line_regex: RegexSearch::new(PYTHON_FILE_LINE_REGEX).unwrap(),
44 }
45 }
46}
47
48pub(super) fn find_from_grid_point<T: EventListener>(
49 term: &Term<T>,
50 point: AlacPoint,
51 regex_searches: &mut RegexSearches,
52) -> Option<(String, bool, Match)> {
53 let grid = term.grid();
54 let link = grid.index(point).hyperlink();
55 let found_word = if link.is_some() {
56 let mut min_index = point;
57 loop {
58 let new_min_index = min_index.sub(term, Boundary::Cursor, 1);
59 if new_min_index == min_index || grid.index(new_min_index).hyperlink() != link {
60 break;
61 } else {
62 min_index = new_min_index
63 }
64 }
65
66 let mut max_index = point;
67 loop {
68 let new_max_index = max_index.add(term, Boundary::Cursor, 1);
69 if new_max_index == max_index || grid.index(new_max_index).hyperlink() != link {
70 break;
71 } else {
72 max_index = new_max_index
73 }
74 }
75
76 let url = link.unwrap().uri().to_owned();
77 let url_match = min_index..=max_index;
78
79 Some((url, true, url_match))
80 } else if let Some(url_match) = regex_match_at(term, point, &mut regex_searches.url_regex) {
81 let url = term.bounds_to_string(*url_match.start(), *url_match.end());
82 Some((url, true, url_match))
83 } else if let Some(python_match) =
84 regex_match_at(term, point, &mut regex_searches.python_file_line_regex)
85 {
86 let matching_line = term.bounds_to_string(*python_match.start(), *python_match.end());
87 python_extract_path_and_line(&matching_line).map(|(file_path, line_number)| {
88 (format!("{file_path}:{line_number}"), false, python_match)
89 })
90 } else if let Some(word_match) = regex_match_at(term, point, &mut regex_searches.word_regex) {
91 let file_path = term.bounds_to_string(*word_match.start(), *word_match.end());
92
93 let (sanitized_match, sanitized_word) = 'sanitize: {
94 let mut word_match = word_match;
95 let mut file_path = file_path;
96
97 if is_path_surrounded_by_common_symbols(&file_path) {
98 word_match = Match::new(
99 word_match.start().add(term, Boundary::Grid, 1),
100 word_match.end().sub(term, Boundary::Grid, 1),
101 );
102 file_path = file_path[1..file_path.len() - 1].to_owned();
103 }
104
105 while file_path.ends_with(':') {
106 file_path.pop();
107 word_match = Match::new(
108 *word_match.start(),
109 word_match.end().sub(term, Boundary::Grid, 1),
110 );
111 }
112 let mut colon_count = 0;
113 for c in file_path.chars() {
114 if c == ':' {
115 colon_count += 1;
116 }
117 }
118 // strip trailing comment after colon in case of
119 // file/at/path.rs:row:column:description or error message
120 // so that the file path is `file/at/path.rs:row:column`
121 if colon_count > 2 {
122 let last_index = file_path.rfind(':').unwrap();
123 let prev_is_digit = last_index > 0
124 && file_path
125 .chars()
126 .nth(last_index - 1)
127 .map_or(false, |c| c.is_ascii_digit());
128 let next_is_digit = last_index < file_path.len() - 1
129 && file_path
130 .chars()
131 .nth(last_index + 1)
132 .map_or(true, |c| c.is_ascii_digit());
133 if prev_is_digit && !next_is_digit {
134 let stripped_len = file_path.len() - last_index;
135 word_match = Match::new(
136 *word_match.start(),
137 word_match.end().sub(term, Boundary::Grid, stripped_len),
138 );
139 file_path = file_path[0..last_index].to_owned();
140 }
141 }
142
143 break 'sanitize (word_match, file_path);
144 };
145
146 Some((sanitized_word, false, sanitized_match))
147 } else {
148 None
149 };
150
151 found_word.map(|(maybe_url_or_path, is_url, word_match)| {
152 if is_url {
153 // Treat "file://" IRIs like file paths to ensure
154 // that line numbers at the end of the path are
155 // handled correctly
156 if let Some(path) = maybe_url_or_path.strip_prefix("file://") {
157 (path.to_string(), false, word_match)
158 } else {
159 (maybe_url_or_path, true, word_match)
160 }
161 } else {
162 (maybe_url_or_path, false, word_match)
163 }
164 })
165}
166
167fn is_path_surrounded_by_common_symbols(path: &str) -> bool {
168 // Avoid detecting `[]` or `()` strings as paths, surrounded by common symbols
169 path.len() > 2
170 // The rest of the brackets and various quotes cannot be matched by the [`WORD_REGEX`] hence not checked for.
171 && (path.starts_with('[') && path.ends_with(']')
172 || path.starts_with('(') && path.ends_with(')'))
173}
174
175/// Based on alacritty/src/display/hint.rs > regex_match_at
176/// Retrieve the match, if the specified point is inside the content matching the regex.
177fn regex_match_at<T>(term: &Term<T>, point: AlacPoint, regex: &mut RegexSearch) -> Option<Match> {
178 visible_regex_match_iter(term, regex).find(|rm| rm.contains(&point))
179}
180
181/// Copied from alacritty/src/display/hint.rs:
182/// Iterate over all visible regex matches.
183fn visible_regex_match_iter<'a, T>(
184 term: &'a Term<T>,
185 regex: &'a mut RegexSearch,
186) -> impl Iterator<Item = Match> + 'a {
187 const MAX_SEARCH_LINES: usize = 100;
188
189 let viewport_start = Line(-(term.grid().display_offset() as i32));
190 let viewport_end = viewport_start + term.bottommost_line();
191 let mut start = term.line_search_left(AlacPoint::new(viewport_start, Column(0)));
192 let mut end = term.line_search_right(AlacPoint::new(viewport_end, Column(0)));
193 start.line = start.line.max(viewport_start - MAX_SEARCH_LINES);
194 end.line = end.line.min(viewport_end + MAX_SEARCH_LINES);
195
196 RegexIter::new(start, end, AlacDirection::Right, term, regex)
197 .skip_while(move |rm| rm.end().line < viewport_start)
198 .take_while(move |rm| rm.start().line <= viewport_end)
199}
200
201#[cfg(test)]
202mod tests {
203 use super::*;
204 use alacritty_terminal::{
205 event::VoidListener,
206 index::{Boundary, Point as AlacPoint},
207 term::{Config, cell::Flags, test::TermSize},
208 vte::ansi::Handler,
209 };
210 use std::{cell::RefCell, ops::RangeInclusive, path::PathBuf};
211 use url::Url;
212 use util::paths::PathWithPosition;
213
214 fn re_test(re: &str, hay: &str, expected: Vec<&str>) {
215 let results: Vec<_> = regex::Regex::new(re)
216 .unwrap()
217 .find_iter(hay)
218 .map(|m| m.as_str())
219 .collect();
220 assert_eq!(results, expected);
221 }
222
223 #[test]
224 fn test_url_regex() {
225 re_test(
226 URL_REGEX,
227 "test http://example.com test mailto:bob@example.com train",
228 vec!["http://example.com", "mailto:bob@example.com"],
229 );
230 }
231
232 #[test]
233 fn test_word_regex() {
234 re_test(
235 WORD_REGEX,
236 "hello, world! \"What\" is this?",
237 vec!["hello", "world", "What", "is", "this"],
238 );
239 }
240
241 #[test]
242 fn test_word_regex_with_linenum() {
243 // filename(line) and filename(line,col) as used in MSBuild output
244 // should be considered a single "word", even though comma is
245 // usually a word separator
246 re_test(WORD_REGEX, "a Main.cs(20) b", vec!["a", "Main.cs(20)", "b"]);
247 re_test(
248 WORD_REGEX,
249 "Main.cs(20,5) Error desc",
250 vec!["Main.cs(20,5)", "Error", "desc"],
251 );
252 // filename:line:col is a popular format for unix tools
253 re_test(
254 WORD_REGEX,
255 "a Main.cs:20:5 b",
256 vec!["a", "Main.cs:20:5", "b"],
257 );
258 // Some tools output "filename:line:col:message", which currently isn't
259 // handled correctly, but might be in the future
260 re_test(
261 WORD_REGEX,
262 "Main.cs:20:5:Error desc",
263 vec!["Main.cs:20:5:Error", "desc"],
264 );
265 }
266
267 #[test]
268 fn test_python_file_line_regex() {
269 re_test(
270 PYTHON_FILE_LINE_REGEX,
271 "hay File \"/zed/bad_py.py\", line 8 stack",
272 vec!["File \"/zed/bad_py.py\", line 8"],
273 );
274 re_test(PYTHON_FILE_LINE_REGEX, "unrelated", vec![]);
275 }
276
277 #[test]
278 fn test_python_file_line() {
279 let inputs: Vec<(&str, Option<(&str, u32)>)> = vec![
280 (
281 "File \"/zed/bad_py.py\", line 8",
282 Some(("/zed/bad_py.py", 8u32)),
283 ),
284 ("File \"path/to/zed/bad_py.py\"", None),
285 ("unrelated", None),
286 ("", None),
287 ];
288 let actual = inputs
289 .iter()
290 .map(|input| python_extract_path_and_line(input.0))
291 .collect::<Vec<_>>();
292 let expected = inputs.iter().map(|(_, output)| *output).collect::<Vec<_>>();
293 assert_eq!(actual, expected);
294 }
295
296 // We use custom columns in many tests to workaround this issue by ensuring a wrapped
297 // line never ends on a wide char:
298 //
299 // <https://github.com/alacritty/alacritty/issues/8586>
300 //
301 // This issue was recently fixed, as soon as we update to a version containing the fix we
302 // can remove all the custom columns from these tests.
303 //
304 macro_rules! test_hyperlink {
305 ($($lines:expr),+; $hyperlink_kind:ident) => { {
306 use crate::terminal_hyperlinks::tests::line_cells_count;
307 use std::cmp;
308
309 let test_lines = vec![$($lines),+];
310 let (total_cells, longest_line_cells) =
311 test_lines.iter().copied()
312 .map(line_cells_count)
313 .fold((0, 0), |state, cells| (state.0 + cells, cmp::max(state.1, cells)));
314
315 test_hyperlink!(
316 // Alacritty has issues with 2 columns, use 3 as the minimum for now.
317 [3, longest_line_cells / 2, longest_line_cells + 1];
318 total_cells;
319 test_lines.iter().copied();
320 $hyperlink_kind
321 )
322 } };
323
324 ($($columns:literal),+; $($lines:expr),+; $hyperlink_kind:ident) => { {
325 use crate::terminal_hyperlinks::tests::line_cells_count;
326
327 let test_lines = vec![$($lines),+];
328 let total_cells = test_lines.iter().copied().map(line_cells_count).sum();
329
330 test_hyperlink!(
331 [ $($columns),+ ]; total_cells; test_lines.iter().copied(); $hyperlink_kind
332 )
333 } };
334
335 ([ $($columns:expr),+ ]; $total_cells:expr; $lines:expr; $hyperlink_kind:ident) => { {
336 use crate::terminal_hyperlinks::tests::{ test_hyperlink, HyperlinkKind };
337
338 let source_location = format!("{}:{}", std::file!(), std::line!());
339 for columns in vec![ $($columns),+] {
340 test_hyperlink(columns, $total_cells, $lines, HyperlinkKind::$hyperlink_kind,
341 &source_location);
342 }
343 } };
344 }
345
346 mod path {
347 /// 👉 := **hovered** on following char
348 ///
349 /// 👈 := **hovered** on wide char spacer of previous full width char
350 ///
351 /// **`‹›`** := expected **hyperlink** match
352 ///
353 /// **`«»`** := expected **path**, **row**, and **column** capture groups
354 ///
355 /// [**`c₀, c₁, …, cₙ;`**]ₒₚₜ := use specified terminal widths of `c₀, c₁, …, cₙ` **columns**
356 /// (defaults to `3, longest_line_cells / 2, longest_line_cells + 1;`)
357 ///
358 macro_rules! test_path {
359 ($($lines:literal),+) => { test_hyperlink!($($lines),+; Path) };
360 ($($columns:literal),+; $($lines:literal),+) => {
361 test_hyperlink!($($columns),+; $($lines),+; Path)
362 };
363 }
364
365 #[test]
366 fn simple() {
367 // Rust paths
368 // Just the path
369 test_path!("‹«/👉test/cool.rs»›");
370 test_path!("‹«/test/cool👉.rs»›");
371
372 // path and line
373 test_path!("‹«/👉test/cool.rs»:«4»›");
374 test_path!("‹«/test/cool.rs»👉:«4»›");
375 test_path!("‹«/test/cool.rs»:«👉4»›");
376 test_path!("‹«/👉test/cool.rs»(«4»)›");
377 test_path!("‹«/test/cool.rs»👉(«4»)›");
378 test_path!("‹«/test/cool.rs»(«👉4»)›");
379 test_path!("‹«/test/cool.rs»(«4»👉)›");
380
381 // path, line, and column
382 test_path!("‹«/👉test/cool.rs»:«4»:«2»›");
383 test_path!("‹«/test/cool.rs»:«4»:«👉2»›");
384 test_path!("‹«/👉test/cool.rs»(«4»,«2»)›");
385 test_path!("‹«/test/cool.rs»(«4»👉,«2»)›");
386
387 // path, line, column, and ':' suffix
388 test_path!("‹«/👉test/cool.rs»:«4»:«2»›:");
389 test_path!("‹«/test/cool.rs»:«4»:«👉2»›:");
390 test_path!("‹«/👉test/cool.rs»(«4»,«2»)›:");
391 test_path!("‹«/test/cool.rs»(«4»,«2»👉)›:");
392
393 // path, line, column, and description
394 test_path!("‹«/test/cool.rs»:«4»:«2»›👉:Error!");
395 test_path!("‹«/test/cool.rs»:«4»:«2»›:👉Error!");
396 test_path!("‹«/test/co👉ol.rs»(«4»,«2»)›:Error!");
397
398 // Cargo output
399 test_path!(" Compiling Cool 👉(‹«/test/Cool»›)");
400 test_path!(" Compiling Cool (‹«/👉test/Cool»›)");
401 test_path!(" Compiling Cool (‹«/test/Cool»›👉)");
402
403 // Python
404 test_path!("‹«awe👉some.py»›");
405
406 test_path!(" ‹F👉ile \"«/awesome.py»\", line «42»›: Wat?");
407 test_path!(" ‹File \"«/awe👉some.py»\", line «42»›: Wat?");
408 test_path!(" ‹File \"«/awesome.py»👉\", line «42»›: Wat?");
409 test_path!(" ‹File \"«/awesome.py»\", line «4👉2»›: Wat?");
410 }
411
412 #[test]
413 fn colons_galore() {
414 test_path!("‹«/test/co👉ol.rs»:«4»›");
415 test_path!("‹«/test/co👉ol.rs»:«4»›:");
416 test_path!("‹«/test/co👉ol.rs»:«4»:«2»›");
417 test_path!("‹«/test/co👉ol.rs»:«4»:«2»›:");
418 test_path!("‹«/test/co👉ol.rs»(«1»)›");
419 test_path!("‹«/test/co👉ol.rs»(«1»)›:");
420 test_path!("‹«/test/co👉ol.rs»(«1»,«618»)›");
421 test_path!("‹«/test/co👉ol.rs»(«1»,«618»)›:");
422 test_path!("‹«/test/co👉ol.rs»::«42»›");
423 test_path!("‹«/test/co👉ol.rs»::«42»›:");
424 test_path!("‹«/test/co👉ol.rs:4:2»(«1»,«618»)›");
425 test_path!("‹«/test/co👉ol.rs»(«1»,«618»)›::");
426 }
427
428 #[test]
429 fn word_wide_chars() {
430 // Rust paths
431 test_path!(4, 6, 12; "‹«/👉例/cool.rs»›");
432 test_path!(4, 6, 12; "‹«/例👈/cool.rs»›");
433 test_path!(4, 8, 16; "‹«/例/cool.rs»:«👉4»›");
434 test_path!(4, 8, 16; "‹«/例/cool.rs»:«4»:«👉2»›");
435
436 // Cargo output
437 test_path!(4, 27, 30; " Compiling Cool (‹«/👉例/Cool»›)");
438 test_path!(4, 27, 30; " Compiling Cool (‹«/例👈/Cool»›)");
439
440 // Python
441 test_path!(4, 11; "‹«👉例wesome.py»›");
442 test_path!(4, 11; "‹«例👈wesome.py»›");
443 test_path!(6, 17, 40; " ‹File \"«/👉例wesome.py»\", line «42»›: Wat?");
444 test_path!(6, 17, 40; " ‹File \"«/例👈wesome.py»\", line «42»›: Wat?");
445 }
446
447 #[test]
448 fn non_word_wide_chars() {
449 // Mojo diagnostic message
450 test_path!(4, 18, 38; " ‹File \"«/awe👉some.🔥»\", line «42»›: Wat?");
451 test_path!(4, 18, 38; " ‹File \"«/awesome👉.🔥»\", line «42»›: Wat?");
452 test_path!(4, 18, 38; " ‹File \"«/awesome.👉🔥»\", line «42»›: Wat?");
453 test_path!(4, 18, 38; " ‹File \"«/awesome.🔥👈»\", line «42»›: Wat?");
454 }
455
456 /// These likely rise to the level of being worth fixing.
457 mod issues {
458 #[test]
459 #[cfg_attr(not(target_os = "windows"), should_panic(expected = "Path = «例»"))]
460 #[cfg_attr(target_os = "windows", should_panic(expected = r#"Path = «C:\\例»"#))]
461 // <https://github.com/alacritty/alacritty/issues/8586>
462 fn issue_alacritty_8586() {
463 // Rust paths
464 test_path!("‹«/👉例/cool.rs»›");
465 test_path!("‹«/例👈/cool.rs»›");
466 test_path!("‹«/例/cool.rs»:«👉4»›");
467 test_path!("‹«/例/cool.rs»:«4»:«👉2»›");
468
469 // Cargo output
470 test_path!(" Compiling Cool (‹«/👉例/Cool»›)");
471 test_path!(" Compiling Cool (‹«/例👈/Cool»›)");
472
473 // Python
474 test_path!("‹«👉例wesome.py»›");
475 test_path!("‹«例👈wesome.py»›");
476 test_path!(" ‹File \"«/👉例wesome.py»\", line «42»›: Wat?");
477 test_path!(" ‹File \"«/例👈wesome.py»\", line «42»›: Wat?");
478 }
479
480 #[test]
481 #[should_panic(expected = "No hyperlink found")]
482 // <https://github.com/zed-industries/zed/issues/12338>
483 fn issue_12338() {
484 // Issue #12338
485 test_path!(".rw-r--r-- 0 staff 05-27 14:03 ‹«test👉、2.txt»›");
486 test_path!(".rw-r--r-- 0 staff 05-27 14:03 ‹«test、👈2.txt»›");
487 test_path!(".rw-r--r-- 0 staff 05-27 14:03 ‹«test👉。3.txt»›");
488 test_path!(".rw-r--r-- 0 staff 05-27 14:03 ‹«test。👈3.txt»›");
489
490 // Rust paths
491 test_path!("‹«/👉🏃/🦀.rs»›");
492 test_path!("‹«/🏃👈/🦀.rs»›");
493 test_path!("‹«/🏃/👉🦀.rs»:«4»›");
494 test_path!("‹«/🏃/🦀👈.rs»:«4»:«2»›");
495
496 // Cargo output
497 test_path!(" Compiling Cool (‹«/👉🏃/Cool»›)");
498 test_path!(" Compiling Cool (‹«/🏃👈/Cool»›)");
499
500 // Python
501 test_path!("‹«👉🏃wesome.py»›");
502 test_path!("‹«🏃👈wesome.py»›");
503 test_path!(" ‹File \"«/👉🏃wesome.py»\", line «42»›: Wat?");
504 test_path!(" ‹File \"«/🏃👈wesome.py»\", line «42»›: Wat?");
505
506 // Mojo
507 test_path!("‹«/awe👉some.🔥»› is some good Mojo!");
508 test_path!("‹«/awesome👉.🔥»› is some good Mojo!");
509 test_path!("‹«/awesome.👉🔥»› is some good Mojo!");
510 test_path!("‹«/awesome.🔥👈»› is some good Mojo!");
511 test_path!(" ‹File \"«/👉🏃wesome.🔥»\", line «42»›: Wat?");
512 test_path!(" ‹File \"«/🏃👈wesome.🔥»\", line «42»›: Wat?");
513 }
514
515 #[test]
516 #[cfg_attr(
517 not(target_os = "windows"),
518 should_panic(
519 expected = "Path = «test/controllers/template_items_controller_test.rb», line = 20, at grid cells (0, 0)..=(17, 1)"
520 )
521 )]
522 #[cfg_attr(
523 target_os = "windows",
524 should_panic(
525 expected = r#"Path = «test\\controllers\\template_items_controller_test.rb», line = 20, at grid cells (0, 0)..=(17, 1)"#
526 )
527 )]
528 // <https://github.com/zed-industries/zed/issues/28194>
529 //
530 // #28194 was closed, but the link includes the description part (":in" here), which
531 // seems wrong...
532 fn issue_28194() {
533 test_path!(
534 "‹«test/c👉ontrollers/template_items_controller_test.rb»:«20»›:in 'block (2 levels) in <class:TemplateItemsControllerTest>'"
535 );
536 test_path!(
537 "‹«test/controllers/template_items_controller_test.rb»:«19»›:i👉n 'block in <class:TemplateItemsControllerTest>'"
538 );
539 }
540 }
541
542 /// Minor issues arguably not important enough to fix/workaround...
543 mod nits {
544 #[test]
545 #[cfg_attr(
546 not(target_os = "windows"),
547 should_panic(expected = "Path = «/test/cool.rs(4»")
548 )]
549 #[cfg_attr(
550 target_os = "windows",
551 should_panic(expected = r#"Path = «C:\\test\\cool.rs(4»"#)
552 )]
553 fn alacritty_bugs_with_two_columns() {
554 test_path!(2; "‹«/👉test/cool.rs»(«4»)›");
555 test_path!(2; "‹«/test/cool.rs»(«👉4»)›");
556 test_path!(2; "‹«/test/cool.rs»(«4»,«👉2»)›");
557
558 // Python
559 test_path!(2; "‹«awe👉some.py»›");
560 }
561
562 #[test]
563 #[cfg_attr(
564 not(target_os = "windows"),
565 should_panic(
566 expected = "Path = «/test/cool.rs», line = 1, at grid cells (0, 0)..=(9, 0)"
567 )
568 )]
569 #[cfg_attr(
570 target_os = "windows",
571 should_panic(
572 expected = r#"Path = «C:\\test\\cool.rs», line = 1, at grid cells (0, 0)..=(9, 2)"#
573 )
574 )]
575 fn invalid_row_column_should_be_part_of_path() {
576 test_path!("‹«/👉test/cool.rs:1:618033988749»›");
577 test_path!("‹«/👉test/cool.rs(1,618033988749)»›");
578 }
579
580 #[test]
581 #[should_panic(expected = "Path = «»")]
582 fn colon_suffix_succeeds_in_finding_an_empty_maybe_path() {
583 test_path!("‹«/test/cool.rs»:«4»:«2»›👉:", "What is this?");
584 test_path!("‹«/test/cool.rs»(«4»,«2»)›👉:", "What is this?");
585 }
586
587 #[test]
588 #[cfg_attr(
589 not(target_os = "windows"),
590 should_panic(expected = "Path = «/test/cool.rs»")
591 )]
592 #[cfg_attr(
593 target_os = "windows",
594 should_panic(expected = r#"Path = «C:\\test\\cool.rs»"#)
595 )]
596 fn many_trailing_colons_should_be_parsed_as_part_of_the_path() {
597 test_path!("‹«/test/cool.rs:::👉:»›");
598 test_path!("‹«/te:st/👉co:ol.r:s:4:2::::::»›");
599 }
600 }
601
602 #[cfg(target_os = "windows")]
603 mod windows {
604 // Lots of fun to be had with long file paths (verbatim) and UNC paths on Windows.
605 // See <https://learn.microsoft.com/en-us/windows/win32/fileio/maximum-file-path-limitation>
606 // See <https://users.rust-lang.org/t/understanding-windows-paths/58583>
607 // See <https://github.com/rust-lang/cargo/issues/13919>
608
609 #[test]
610 fn unc() {
611 test_path!(r#"‹«\\server\share\👉test\cool.rs»›"#);
612 test_path!(r#"‹«\\server\share\test\cool👉.rs»›"#);
613 }
614
615 mod issues {
616 #[test]
617 #[should_panic(
618 expected = r#"Path = «C:\\test\\cool.rs», at grid cells (0, 0)..=(6, 0)"#
619 )]
620 fn issue_verbatim() {
621 test_path!(r#"‹«\\?\C:\👉test\cool.rs»›"#);
622 test_path!(r#"‹«\\?\C:\test\cool👉.rs»›"#);
623 }
624
625 #[test]
626 #[should_panic(
627 expected = r#"Path = «\\\\server\\share\\test\\cool.rs», at grid cells (0, 0)..=(10, 2)"#
628 )]
629 fn issue_verbatim_unc() {
630 test_path!(r#"‹«\\?\UNC\server\share\👉test\cool.rs»›"#);
631 test_path!(r#"‹«\\?\UNC\server\share\test\cool👉.rs»›"#);
632 }
633 }
634 }
635 }
636
637 mod file_iri {
638 // File IRIs have a ton of use cases, most of which we currently do not support. A few of
639 // those cases are documented here as tests which are expected to fail.
640 // See https://en.wikipedia.org/wiki/File_URI_scheme
641
642 /// [**`c₀, c₁, …, cₙ;`**]ₒₚₜ := use specified terminal widths of `c₀, c₁, …, cₙ` **columns**
643 /// (defaults to `3, longest_line_cells / 2, longest_line_cells + 1;`)
644 ///
645 macro_rules! test_file_iri {
646 ($file_iri:literal) => { { test_hyperlink!(concat!("‹«👉", $file_iri, "»›"); FileIri) } };
647 ($($columns:literal),+; $file_iri:literal) => { {
648 test_hyperlink!($($columns),+; concat!("‹«👉", $file_iri, "»›"); FileIri)
649 } };
650 }
651
652 #[cfg(not(target_os = "windows"))]
653 #[test]
654 fn absolute_file_iri() {
655 test_file_iri!("file:///test/cool/index.rs");
656 test_file_iri!("file:///test/cool/");
657 }
658
659 mod issues {
660 #[cfg(not(target_os = "windows"))]
661 #[test]
662 #[should_panic(expected = "Path = «/test/Ῥόδος/», at grid cells (0, 0)..=(15, 1)")]
663 fn issue_file_iri_with_percent_encoded_characters() {
664 // Non-space characters
665 // file:///test/Ῥόδος/
666 test_file_iri!("file:///test/%E1%BF%AC%CF%8C%CE%B4%CE%BF%CF%82/"); // URI
667
668 // Spaces
669 test_file_iri!("file:///te%20st/co%20ol/index.rs");
670 test_file_iri!("file:///te%20st/co%20ol/");
671 }
672 }
673
674 #[cfg(target_os = "windows")]
675 mod windows {
676 mod issues {
677 // The test uses Url::to_file_path(), but it seems that the Url crate doesn't
678 // support relative file IRIs.
679 #[test]
680 #[should_panic(
681 expected = r#"Failed to interpret file IRI `file:/test/cool/index.rs` as a path"#
682 )]
683 fn issue_relative_file_iri() {
684 test_file_iri!("file:/test/cool/index.rs");
685 test_file_iri!("file:/test/cool/");
686 }
687
688 // See https://en.wikipedia.org/wiki/File_URI_scheme
689 #[test]
690 #[should_panic(
691 expected = r#"Path = «C:\\test\\cool\\index.rs», at grid cells (0, 0)..=(9, 1)"#
692 )]
693 fn issue_absolute_file_iri() {
694 test_file_iri!("file:///C:/test/cool/index.rs");
695 test_file_iri!("file:///C:/test/cool/");
696 }
697
698 #[test]
699 #[should_panic(
700 expected = r#"Path = «C:\\test\\Ῥόδος\\», at grid cells (0, 0)..=(16, 1)"#
701 )]
702 fn issue_file_iri_with_percent_encoded_characters() {
703 // Non-space characters
704 // file:///test/Ῥόδος/
705 test_file_iri!("file:///C:/test/%E1%BF%AC%CF%8C%CE%B4%CE%BF%CF%82/"); // URI
706
707 // Spaces
708 test_file_iri!("file:///C:/te%20st/co%20ol/index.rs");
709 test_file_iri!("file:///C:/te%20st/co%20ol/");
710 }
711 }
712 }
713 }
714
715 mod iri {
716 /// [**`c₀, c₁, …, cₙ;`**]ₒₚₜ := use specified terminal widths of `c₀, c₁, …, cₙ` **columns**
717 /// (defaults to `3, longest_line_cells / 2, longest_line_cells + 1;`)
718 ///
719 macro_rules! test_iri {
720 ($iri:literal) => { { test_hyperlink!(concat!("‹«👉", $iri, "»›"); Iri) } };
721 ($($columns:literal),+; $iri:literal) => { {
722 test_hyperlink!($($columns),+; concat!("‹«👉", $iri, "»›"); Iri)
723 } };
724 }
725
726 #[test]
727 fn simple() {
728 // In the order they appear in URL_REGEX, except 'file://' which is treated as a path
729 test_iri!("ipfs://test/cool.ipfs");
730 test_iri!("ipns://test/cool.ipns");
731 test_iri!("magnet://test/cool.git");
732 test_iri!("mailto:someone@somewhere.here");
733 test_iri!("gemini://somewhere.here");
734 test_iri!("gopher://somewhere.here");
735 test_iri!("http://test/cool/index.html");
736 test_iri!("http://10.10.10.10:1111/cool.html");
737 test_iri!("http://test/cool/index.html?amazing=1");
738 test_iri!("http://test/cool/index.html#right%20here");
739 test_iri!("http://test/cool/index.html?amazing=1#right%20here");
740 test_iri!("https://test/cool/index.html");
741 test_iri!("https://10.10.10.10:1111/cool.html");
742 test_iri!("https://test/cool/index.html?amazing=1");
743 test_iri!("https://test/cool/index.html#right%20here");
744 test_iri!("https://test/cool/index.html?amazing=1#right%20here");
745 test_iri!("news://test/cool.news");
746 test_iri!("git://test/cool.git");
747 test_iri!("ssh://user@somewhere.over.here:12345/test/cool.git");
748 test_iri!("ftp://test/cool.ftp");
749 }
750
751 #[test]
752 fn wide_chars() {
753 // In the order they appear in URL_REGEX, except 'file://' which is treated as a path
754 test_iri!(4, 20; "ipfs://例🏃🦀/cool.ipfs");
755 test_iri!(4, 20; "ipns://例🏃🦀/cool.ipns");
756 test_iri!(6, 20; "magnet://例🏃🦀/cool.git");
757 test_iri!(4, 20; "mailto:someone@somewhere.here");
758 test_iri!(4, 20; "gemini://somewhere.here");
759 test_iri!(4, 20; "gopher://somewhere.here");
760 test_iri!(4, 20; "http://例🏃🦀/cool/index.html");
761 test_iri!(4, 20; "http://10.10.10.10:1111/cool.html");
762 test_iri!(4, 20; "http://例🏃🦀/cool/index.html?amazing=1");
763 test_iri!(4, 20; "http://例🏃🦀/cool/index.html#right%20here");
764 test_iri!(4, 20; "http://例🏃🦀/cool/index.html?amazing=1#right%20here");
765 test_iri!(4, 20; "https://例🏃🦀/cool/index.html");
766 test_iri!(4, 20; "https://10.10.10.10:1111/cool.html");
767 test_iri!(4, 20; "https://例🏃🦀/cool/index.html?amazing=1");
768 test_iri!(4, 20; "https://例🏃🦀/cool/index.html#right%20here");
769 test_iri!(4, 20; "https://例🏃🦀/cool/index.html?amazing=1#right%20here");
770 test_iri!(4, 20; "news://例🏃🦀/cool.news");
771 test_iri!(5, 20; "git://例/cool.git");
772 test_iri!(5, 20; "ssh://user@somewhere.over.here:12345/例🏃🦀/cool.git");
773 test_iri!(7, 20; "ftp://例🏃🦀/cool.ftp");
774 }
775
776 // There are likely more tests needed for IRI vs URI
777 #[test]
778 fn iris() {
779 // These refer to the same location, see example here:
780 // <https://en.wikipedia.org/wiki/Internationalized_Resource_Identifier#Compatibility>
781 test_iri!("https://en.wiktionary.org/wiki/Ῥόδος"); // IRI
782 test_iri!("https://en.wiktionary.org/wiki/%E1%BF%AC%CF%8C%CE%B4%CE%BF%CF%82"); // URI
783 }
784
785 #[test]
786 #[should_panic(expected = "Expected a path, but was a iri")]
787 fn file_is_a_path() {
788 test_iri!("file://test/cool/index.rs");
789 }
790 }
791
792 #[derive(Debug, PartialEq)]
793 enum HyperlinkKind {
794 FileIri,
795 Iri,
796 Path,
797 }
798
799 struct ExpectedHyperlink {
800 hovered_grid_point: AlacPoint,
801 hovered_char: char,
802 hyperlink_kind: HyperlinkKind,
803 iri_or_path: String,
804 row: Option<u32>,
805 column: Option<u32>,
806 hyperlink_match: RangeInclusive<AlacPoint>,
807 }
808
809 /// Converts to Windows style paths on Windows, like path!(), but at runtime for improved test
810 /// readability.
811 fn build_term_from_test_lines<'a>(
812 hyperlink_kind: HyperlinkKind,
813 term_size: TermSize,
814 test_lines: impl Iterator<Item = &'a str>,
815 ) -> (Term<VoidListener>, ExpectedHyperlink) {
816 #[derive(Default, Eq, PartialEq)]
817 enum HoveredState {
818 #[default]
819 HoveredScan,
820 HoveredNextChar,
821 Done,
822 }
823
824 #[derive(Default, Eq, PartialEq)]
825 enum MatchState {
826 #[default]
827 MatchScan,
828 MatchNextChar,
829 Match(AlacPoint),
830 Done,
831 }
832
833 #[derive(Default, Eq, PartialEq)]
834 enum CapturesState {
835 #[default]
836 PathScan,
837 PathNextChar,
838 Path(AlacPoint),
839 RowScan,
840 Row(String),
841 ColumnScan,
842 Column(String),
843 Done,
844 }
845
846 fn prev_input_point_from_term(term: &Term<VoidListener>) -> AlacPoint {
847 let grid = term.grid();
848 let cursor = &grid.cursor;
849 let mut point = cursor.point;
850
851 if !cursor.input_needs_wrap {
852 point.column -= 1;
853 }
854
855 if grid.index(point).flags.contains(Flags::WIDE_CHAR_SPACER) {
856 point.column -= 1;
857 }
858
859 point
860 }
861
862 let mut hovered_grid_point: Option<AlacPoint> = None;
863 let mut hyperlink_match = AlacPoint::default()..=AlacPoint::default();
864 let mut iri_or_path = String::default();
865 let mut row = None;
866 let mut column = None;
867 let mut prev_input_point = AlacPoint::default();
868 let mut hovered_state = HoveredState::default();
869 let mut match_state = MatchState::default();
870 let mut captures_state = CapturesState::default();
871 let mut term = Term::new(Config::default(), &term_size, VoidListener);
872
873 for text in test_lines {
874 let chars: Box<dyn Iterator<Item = char>> =
875 if cfg!(windows) && hyperlink_kind == HyperlinkKind::Path {
876 Box::new(text.chars().map(|c| if c == '/' { '\\' } else { c })) as _
877 } else {
878 Box::new(text.chars()) as _
879 };
880 let mut chars = chars.peekable();
881 while let Some(c) = chars.next() {
882 match c {
883 '👉' => {
884 hovered_state = HoveredState::HoveredNextChar;
885 }
886 '👈' => {
887 hovered_grid_point = Some(prev_input_point.add(&term, Boundary::Grid, 1));
888 }
889 '«' | '»' => {
890 captures_state = match captures_state {
891 CapturesState::PathScan => CapturesState::PathNextChar,
892 CapturesState::PathNextChar => {
893 panic!("Should have been handled by char input")
894 }
895 CapturesState::Path(start_point) => {
896 iri_or_path = term.bounds_to_string(start_point, prev_input_point);
897 CapturesState::RowScan
898 }
899 CapturesState::RowScan => CapturesState::Row(String::new()),
900 CapturesState::Row(number) => {
901 row = Some(number.parse::<u32>().unwrap());
902 CapturesState::ColumnScan
903 }
904 CapturesState::ColumnScan => CapturesState::Column(String::new()),
905 CapturesState::Column(number) => {
906 column = Some(number.parse::<u32>().unwrap());
907 CapturesState::Done
908 }
909 CapturesState::Done => {
910 panic!("Extra '«', '»'")
911 }
912 }
913 }
914 '‹' | '›' => {
915 match_state = match match_state {
916 MatchState::MatchScan => MatchState::MatchNextChar,
917 MatchState::MatchNextChar => {
918 panic!("Should have been handled by char input")
919 }
920 MatchState::Match(start_point) => {
921 hyperlink_match = start_point..=prev_input_point;
922 MatchState::Done
923 }
924 MatchState::Done => {
925 panic!("Extra '‹', '›'")
926 }
927 }
928 }
929 _ => {
930 if let CapturesState::Row(number) | CapturesState::Column(number) =
931 &mut captures_state
932 {
933 number.push(c)
934 }
935
936 let is_windows_abs_path_start = captures_state
937 == CapturesState::PathNextChar
938 && cfg!(windows)
939 && hyperlink_kind == HyperlinkKind::Path
940 && c == '\\'
941 && chars.peek().is_some_and(|c| *c != '\\');
942
943 if is_windows_abs_path_start {
944 // Convert Unix abs path start into Windows abs path start so that the
945 // same test can be used for both OSes.
946 term.input('C');
947 prev_input_point = prev_input_point_from_term(&term);
948 term.input(':');
949 term.input(c);
950 } else {
951 term.input(c);
952 prev_input_point = prev_input_point_from_term(&term);
953 }
954
955 if hovered_state == HoveredState::HoveredNextChar {
956 hovered_grid_point = Some(prev_input_point);
957 hovered_state = HoveredState::Done;
958 }
959 if captures_state == CapturesState::PathNextChar {
960 captures_state = CapturesState::Path(prev_input_point);
961 }
962 if match_state == MatchState::MatchNextChar {
963 match_state = MatchState::Match(prev_input_point);
964 }
965 }
966 }
967 }
968 term.move_down_and_cr(1);
969 }
970
971 if hyperlink_kind == HyperlinkKind::FileIri {
972 let Ok(url) = Url::parse(&iri_or_path) else {
973 panic!("Failed to parse file IRI `{iri_or_path}`");
974 };
975 let Ok(path) = url.to_file_path() else {
976 panic!("Failed to interpret file IRI `{iri_or_path}` as a path");
977 };
978 iri_or_path = path.to_string_lossy().to_string();
979 }
980
981 if cfg!(windows) {
982 // Handle verbatim and UNC paths for Windows
983 if let Some(stripped) = iri_or_path.strip_prefix(r#"\\?\UNC\"#) {
984 iri_or_path = format!(r#"\\{stripped}"#);
985 } else if let Some(stripped) = iri_or_path.strip_prefix(r#"\\?\"#) {
986 iri_or_path = stripped.to_string();
987 }
988 }
989
990 let hovered_grid_point = hovered_grid_point.expect("Missing hovered point (👉 or 👈)");
991 let hovered_char = term.grid().index(hovered_grid_point).c;
992 (
993 term,
994 ExpectedHyperlink {
995 hovered_grid_point,
996 hovered_char,
997 hyperlink_kind,
998 iri_or_path,
999 row,
1000 column,
1001 hyperlink_match,
1002 },
1003 )
1004 }
1005
1006 fn line_cells_count(line: &str) -> usize {
1007 // This avoids taking a dependency on the unicode-width crate
1008 fn width(c: char) -> usize {
1009 match c {
1010 // Fullwidth unicode characters used in tests
1011 '例' | '🏃' | '🦀' | '🔥' => 2,
1012 _ => 1,
1013 }
1014 }
1015 const CONTROL_CHARS: &str = "‹«👉👈»›";
1016 line.chars()
1017 .filter(|c| !CONTROL_CHARS.contains(*c))
1018 .map(width)
1019 .sum::<usize>()
1020 }
1021
1022 struct CheckHyperlinkMatch<'a> {
1023 term: &'a Term<VoidListener>,
1024 expected_hyperlink: &'a ExpectedHyperlink,
1025 source_location: &'a str,
1026 }
1027
1028 impl<'a> CheckHyperlinkMatch<'a> {
1029 fn new(
1030 term: &'a Term<VoidListener>,
1031 expected_hyperlink: &'a ExpectedHyperlink,
1032 source_location: &'a str,
1033 ) -> Self {
1034 Self {
1035 term,
1036 expected_hyperlink,
1037 source_location,
1038 }
1039 }
1040
1041 fn check_path_with_position_and_match(
1042 &self,
1043 path_with_position: PathWithPosition,
1044 hyperlink_match: &Match,
1045 ) {
1046 let format_path_with_position_and_match =
1047 |path_with_position: &PathWithPosition, hyperlink_match: &Match| {
1048 let mut result =
1049 format!("Path = «{}»", &path_with_position.path.to_string_lossy());
1050 if let Some(row) = path_with_position.row {
1051 result += &format!(", line = {row}");
1052 if let Some(column) = path_with_position.column {
1053 result += &format!(", column = {column}");
1054 }
1055 }
1056
1057 result += &format!(
1058 ", at grid cells {}",
1059 Self::format_hyperlink_match(hyperlink_match)
1060 );
1061 result
1062 };
1063
1064 assert_ne!(
1065 self.expected_hyperlink.hyperlink_kind,
1066 HyperlinkKind::Iri,
1067 "\n at {}\nExpected a path, but was a iri:\n{}",
1068 self.source_location,
1069 self.format_renderable_content()
1070 );
1071
1072 assert_eq!(
1073 format_path_with_position_and_match(
1074 &PathWithPosition {
1075 path: PathBuf::from(self.expected_hyperlink.iri_or_path.clone()),
1076 row: self.expected_hyperlink.row,
1077 column: self.expected_hyperlink.column
1078 },
1079 &self.expected_hyperlink.hyperlink_match
1080 ),
1081 format_path_with_position_and_match(&path_with_position, hyperlink_match),
1082 "\n at {}:\n{}",
1083 self.source_location,
1084 self.format_renderable_content()
1085 );
1086 }
1087
1088 fn check_iri_and_match(&self, iri: String, hyperlink_match: &Match) {
1089 let format_iri_and_match = |iri: &String, hyperlink_match: &Match| {
1090 format!(
1091 "Url = «{iri}», at grid cells {}",
1092 Self::format_hyperlink_match(hyperlink_match)
1093 )
1094 };
1095
1096 assert_eq!(
1097 self.expected_hyperlink.hyperlink_kind,
1098 HyperlinkKind::Iri,
1099 "\n at {}\nExpected a iri, but was a path:\n{}",
1100 self.source_location,
1101 self.format_renderable_content()
1102 );
1103
1104 assert_eq!(
1105 format_iri_and_match(
1106 &self.expected_hyperlink.iri_or_path,
1107 &self.expected_hyperlink.hyperlink_match
1108 ),
1109 format_iri_and_match(&iri, hyperlink_match),
1110 "\n at {}:\n{}",
1111 self.source_location,
1112 self.format_renderable_content()
1113 );
1114 }
1115
1116 fn format_hyperlink_match(hyperlink_match: &Match) -> String {
1117 format!(
1118 "({}, {})..=({}, {})",
1119 hyperlink_match.start().line.0,
1120 hyperlink_match.start().column.0,
1121 hyperlink_match.end().line.0,
1122 hyperlink_match.end().column.0
1123 )
1124 }
1125
1126 fn format_renderable_content(&self) -> String {
1127 let mut result = format!("\nHovered on '{}'\n", self.expected_hyperlink.hovered_char);
1128
1129 let mut first_header_row = String::new();
1130 let mut second_header_row = String::new();
1131 let mut marker_header_row = String::new();
1132 for index in 0..self.term.columns() {
1133 let remainder = index % 10;
1134 first_header_row.push_str(
1135 &(index > 0 && remainder == 0)
1136 .then_some((index / 10).to_string())
1137 .unwrap_or(" ".into()),
1138 );
1139 second_header_row += &remainder.to_string();
1140 if index == self.expected_hyperlink.hovered_grid_point.column.0 {
1141 marker_header_row.push('↓');
1142 } else {
1143 marker_header_row.push(' ');
1144 }
1145 }
1146
1147 result += &format!("\n [{}]\n", first_header_row);
1148 result += &format!(" [{}]\n", second_header_row);
1149 result += &format!(" {}", marker_header_row);
1150
1151 let spacers: Flags = Flags::LEADING_WIDE_CHAR_SPACER | Flags::WIDE_CHAR_SPACER;
1152 for cell in self
1153 .term
1154 .renderable_content()
1155 .display_iter
1156 .filter(|cell| !cell.flags.intersects(spacers))
1157 {
1158 if cell.point.column.0 == 0 {
1159 let prefix =
1160 if cell.point.line == self.expected_hyperlink.hovered_grid_point.line {
1161 '→'
1162 } else {
1163 ' '
1164 };
1165 result += &format!("\n{prefix}[{:>3}] ", cell.point.line.to_string());
1166 }
1167
1168 result.push(cell.c);
1169 }
1170
1171 result
1172 }
1173 }
1174
1175 fn test_hyperlink<'a>(
1176 columns: usize,
1177 total_cells: usize,
1178 test_lines: impl Iterator<Item = &'a str>,
1179 hyperlink_kind: HyperlinkKind,
1180 source_location: &str,
1181 ) {
1182 thread_local! {
1183 static TEST_REGEX_SEARCHES: RefCell<RegexSearches> = RefCell::new(RegexSearches::new());
1184 }
1185
1186 let term_size = TermSize::new(columns, total_cells / columns + 2);
1187 let (term, expected_hyperlink) =
1188 build_term_from_test_lines(hyperlink_kind, term_size, test_lines);
1189 let hyperlink_found = TEST_REGEX_SEARCHES.with(|regex_searches| {
1190 find_from_grid_point(
1191 &term,
1192 expected_hyperlink.hovered_grid_point,
1193 &mut regex_searches.borrow_mut(),
1194 )
1195 });
1196 let check_hyperlink_match =
1197 CheckHyperlinkMatch::new(&term, &expected_hyperlink, source_location);
1198 match hyperlink_found {
1199 Some((hyperlink_word, false, hyperlink_match)) => {
1200 check_hyperlink_match.check_path_with_position_and_match(
1201 PathWithPosition::parse_str(&hyperlink_word),
1202 &hyperlink_match,
1203 );
1204 }
1205 Some((hyperlink_word, true, hyperlink_match)) => {
1206 check_hyperlink_match.check_iri_and_match(hyperlink_word, &hyperlink_match);
1207 }
1208 _ => {
1209 assert!(
1210 false,
1211 "No hyperlink found\n at {source_location}:\n{}",
1212 check_hyperlink_match.format_renderable_content()
1213 )
1214 }
1215 }
1216 }
1217}