Show More
@@ -1,391 +1,377 b'' | |||
|
1 | 1 | // filepatterns.rs |
|
2 | 2 | // |
|
3 | 3 | // Copyright 2019 Raphaël Gomès <rgomes@octobus.net> |
|
4 | 4 | // |
|
5 | 5 | // This software may be used and distributed according to the terms of the |
|
6 | 6 | // GNU General Public License version 2 or any later version. |
|
7 | 7 | |
|
8 | 8 | //! Handling of Mercurial-specific patterns. |
|
9 | 9 | |
|
10 | 10 | use crate::{ |
|
11 | 11 | utils::{files::get_path_from_bytes, SliceExt}, |
|
12 | 12 | LineNumber, PatternError, PatternFileError, |
|
13 | 13 | }; |
|
14 | 14 | use lazy_static::lazy_static; |
|
15 | 15 | use regex::bytes::{NoExpand, Regex}; |
|
16 | 16 | use std::collections::HashMap; |
|
17 | 17 | use std::fs::File; |
|
18 | 18 | use std::io::Read; |
|
19 | 19 | use std::vec::Vec; |
|
20 | 20 | |
|
21 | 21 | lazy_static! { |
|
22 | 22 | static ref RE_ESCAPE: Vec<Vec<u8>> = { |
|
23 | 23 | let mut v: Vec<Vec<u8>> = (0..=255).map(|byte| vec![byte]).collect(); |
|
24 | 24 | let to_escape = b"()[]{}?*+-|^$\\.&~# \t\n\r\x0b\x0c"; |
|
25 | 25 | for byte in to_escape { |
|
26 | 26 | v[*byte as usize].insert(0, b'\\'); |
|
27 | 27 | } |
|
28 | 28 | v |
|
29 | 29 | }; |
|
30 | 30 | } |
|
31 | 31 | |
|
32 | 32 | /// These are matched in order |
|
33 | 33 | const GLOB_REPLACEMENTS: &[(&[u8], &[u8])] = |
|
34 | 34 | &[(b"*/", b"(?:.*/)?"), (b"*", b".*"), (b"", b"[^/]*")]; |
|
35 | 35 | |
|
36 | 36 | #[derive(Debug, Copy, Clone, PartialEq, Eq)] |
|
37 | 37 | pub enum PatternSyntax { |
|
38 | 38 | Regexp, |
|
39 | 39 | /// Glob that matches at the front of the path |
|
40 | 40 | RootGlob, |
|
41 | 41 | /// Glob that matches at any suffix of the path (still anchored at |
|
42 | 42 | /// slashes) |
|
43 | 43 | Glob, |
|
44 | 44 | Path, |
|
45 | 45 | RelPath, |
|
46 | 46 | RelGlob, |
|
47 | 47 | RelRegexp, |
|
48 | 48 | RootFiles, |
|
49 | 49 | } |
|
50 | 50 | |
|
51 | 51 | /// Transforms a glob pattern into a regex |
|
52 | 52 | fn glob_to_re(pat: &[u8]) -> Vec<u8> { |
|
53 | 53 | let mut input = pat; |
|
54 | 54 | let mut res: Vec<u8> = vec![]; |
|
55 | 55 | let mut group_depth = 0; |
|
56 | 56 | |
|
57 | 57 | while let Some((c, rest)) = input.split_first() { |
|
58 | 58 | input = rest; |
|
59 | 59 | |
|
60 | 60 | match c { |
|
61 | 61 | b'*' => { |
|
62 | 62 | for (source, repl) in GLOB_REPLACEMENTS { |
|
63 | 63 | if let Some(rest) = input.drop_prefix(source) { |
|
64 | 64 | input = rest; |
|
65 | 65 | res.extend(*repl); |
|
66 | 66 | break; |
|
67 | 67 | } |
|
68 | 68 | } |
|
69 | 69 | } |
|
70 | 70 | b'?' => res.extend(b"."), |
|
71 | 71 | b'[' => { |
|
72 | 72 | match input.iter().skip(1).position(|b| *b == b']') { |
|
73 | 73 | None => res.extend(b"\\["), |
|
74 | 74 | Some(end) => { |
|
75 | 75 | // Account for the one we skipped |
|
76 | 76 | let end = end + 1; |
|
77 | 77 | |
|
78 | 78 | res.extend(b"["); |
|
79 | 79 | |
|
80 | 80 | for (i, b) in input[..end].iter().enumerate() { |
|
81 | 81 | if *b == b'!' && i == 0 { |
|
82 | 82 | res.extend(b"^") |
|
83 | 83 | } else if *b == b'^' && i == 0 { |
|
84 | 84 | res.extend(b"\\^") |
|
85 | 85 | } else if *b == b'\\' { |
|
86 | 86 | res.extend(b"\\\\") |
|
87 | 87 | } else { |
|
88 | 88 | res.push(*b) |
|
89 | 89 | } |
|
90 | 90 | } |
|
91 | 91 | res.extend(b"]"); |
|
92 | 92 | input = &input[end + 1..]; |
|
93 | 93 | } |
|
94 | 94 | } |
|
95 | 95 | } |
|
96 | 96 | b'{' => { |
|
97 | 97 | group_depth += 1; |
|
98 | 98 | res.extend(b"(?:") |
|
99 | 99 | } |
|
100 | 100 | b'}' if group_depth > 0 => { |
|
101 | 101 | group_depth -= 1; |
|
102 | 102 | res.extend(b")"); |
|
103 | 103 | } |
|
104 | 104 | b',' if group_depth > 0 => res.extend(b"|"), |
|
105 | 105 | b'\\' => { |
|
106 | 106 | let c = { |
|
107 | 107 | if let Some((c, rest)) = input.split_first() { |
|
108 | 108 | input = rest; |
|
109 | 109 | c |
|
110 | 110 | } else { |
|
111 | 111 | c |
|
112 | 112 | } |
|
113 | 113 | }; |
|
114 | 114 | res.extend(&RE_ESCAPE[*c as usize]) |
|
115 | 115 | } |
|
116 | 116 | _ => res.extend(&RE_ESCAPE[*c as usize]), |
|
117 | 117 | } |
|
118 | 118 | } |
|
119 | 119 | res |
|
120 | 120 | } |
|
121 | 121 | |
|
122 | 122 | fn escape_pattern(pattern: &[u8]) -> Vec<u8> { |
|
123 | 123 | pattern |
|
124 | 124 | .iter() |
|
125 | 125 | .flat_map(|c| RE_ESCAPE[*c as usize].clone()) |
|
126 | 126 | .collect() |
|
127 | 127 | } |
|
128 | 128 | |
|
129 | 129 | fn parse_pattern_syntax(kind: &[u8]) -> Result<PatternSyntax, PatternError> { |
|
130 | 130 | match kind { |
|
131 | 131 | b"re" => Ok(PatternSyntax::Regexp), |
|
132 | 132 | b"path" => Ok(PatternSyntax::Path), |
|
133 | 133 | b"relpath" => Ok(PatternSyntax::RelPath), |
|
134 | 134 | b"rootfilesin" => Ok(PatternSyntax::RootFiles), |
|
135 | 135 | b"relglob" => Ok(PatternSyntax::RelGlob), |
|
136 | 136 | b"relre" => Ok(PatternSyntax::RelRegexp), |
|
137 | 137 | b"glob" => Ok(PatternSyntax::Glob), |
|
138 | 138 | b"rootglob" => Ok(PatternSyntax::RootGlob), |
|
139 | 139 | _ => Err(PatternError::UnsupportedSyntax( |
|
140 | 140 | String::from_utf8_lossy(kind).to_string(), |
|
141 | 141 | )), |
|
142 | 142 | } |
|
143 | 143 | } |
|
144 | 144 | |
|
145 | 145 | /// Builds the regex that corresponds to the given pattern. |
|
146 | 146 | /// If within a `syntax: regexp` context, returns the pattern, |
|
147 | 147 | /// otherwise, returns the corresponding regex. |
|
148 | 148 | fn _build_single_regex( |
|
149 | 149 | syntax: PatternSyntax, |
|
150 | 150 | pattern: &[u8], |
|
151 | 151 | globsuffix: &[u8], |
|
152 | 152 | ) -> Vec<u8> { |
|
153 | 153 | if pattern.is_empty() { |
|
154 | 154 | return vec![]; |
|
155 | 155 | } |
|
156 | 156 | match syntax { |
|
157 | 157 | PatternSyntax::Regexp => pattern.to_owned(), |
|
158 | 158 | PatternSyntax::RelRegexp => { |
|
159 | 159 | if pattern[0] == b'^' { |
|
160 | 160 | return pattern.to_owned(); |
|
161 | 161 | } |
|
162 | let mut res = b".*".to_vec(); | |
|
163 | res.extend(pattern); | |
|
164 | res | |
|
162 | [b".*", pattern].concat() | |
|
165 | 163 | } |
|
166 | 164 | PatternSyntax::Path | PatternSyntax::RelPath => { |
|
167 | 165 | if pattern == b"." { |
|
168 | 166 | return vec![]; |
|
169 | 167 | } |
|
170 | let mut pattern = escape_pattern(pattern); | |
|
171 | pattern.extend(b"(?:/|$)"); | |
|
172 | pattern | |
|
168 | [escape_pattern(pattern).as_slice(), b"(?:/|$)"].concat() | |
|
173 | 169 | } |
|
174 | 170 | PatternSyntax::RootFiles => { |
|
175 | 171 | let mut res = if pattern == b"." { |
|
176 | 172 | vec![] |
|
177 | 173 | } else { |
|
178 | 174 | // Pattern is a directory name. |
|
179 | let mut as_vec: Vec<u8> = escape_pattern(pattern); | |
|
180 | as_vec.push(b'/'); | |
|
181 | as_vec | |
|
175 | [escape_pattern(pattern).as_slice(), b"/"].concat() | |
|
182 | 176 | }; |
|
183 | 177 | |
|
184 | 178 | // Anything after the pattern must be a non-directory. |
|
185 | 179 | res.extend(b"[^/]+$"); |
|
186 | 180 | res |
|
187 | 181 | } |
|
188 | 182 | PatternSyntax::RelGlob => { |
|
189 | let mut res: Vec<u8> = vec![]; | |
|
190 | 183 | let glob_re = glob_to_re(pattern); |
|
191 | 184 | if let Some(rest) = glob_re.drop_prefix(b"[^/]*") { |
|
192 | res.extend(b".*"); | |
|
193 | res.extend(rest); | |
|
185 | [b".*", rest, globsuffix].concat() | |
|
194 | 186 | } else { |
|
195 | res.extend(b"(?:|.*/)"); | |
|
196 | res.extend(glob_re); | |
|
187 | [b"(?:|.*/)", glob_re.as_slice(), globsuffix].concat() | |
|
197 | 188 | } |
|
198 | res.extend(globsuffix.iter()); | |
|
199 | res | |
|
200 | 189 | } |
|
201 | 190 | PatternSyntax::Glob | PatternSyntax::RootGlob => { |
|
202 | let mut res: Vec<u8> = vec![]; | |
|
203 | res.extend(glob_to_re(pattern)); | |
|
204 | res.extend(globsuffix.iter()); | |
|
205 | res | |
|
191 | [glob_to_re(pattern).as_slice(), globsuffix].concat() | |
|
206 | 192 | } |
|
207 | 193 | } |
|
208 | 194 | } |
|
209 | 195 | |
|
210 | 196 | const GLOB_SPECIAL_CHARACTERS: [u8; 7] = |
|
211 | 197 | [b'*', b'?', b'[', b']', b'{', b'}', b'\\']; |
|
212 | 198 | |
|
213 | 199 | /// Wrapper function to `_build_single_regex` that short-circuits 'exact' globs |
|
214 | 200 | /// that don't need to be transformed into a regex. |
|
215 | 201 | pub fn build_single_regex( |
|
216 | 202 | kind: &[u8], |
|
217 | 203 | pat: &[u8], |
|
218 | 204 | globsuffix: &[u8], |
|
219 | 205 | ) -> Result<Vec<u8>, PatternError> { |
|
220 | 206 | let enum_kind = parse_pattern_syntax(kind)?; |
|
221 | 207 | if enum_kind == PatternSyntax::RootGlob |
|
222 | 208 | && !pat.iter().any(|b| GLOB_SPECIAL_CHARACTERS.contains(b)) |
|
223 | 209 | { |
|
224 | 210 | let mut escaped = escape_pattern(pat); |
|
225 | 211 | escaped.extend(b"(?:/|$)"); |
|
226 | 212 | Ok(escaped) |
|
227 | 213 | } else { |
|
228 | 214 | Ok(_build_single_regex(enum_kind, pat, globsuffix)) |
|
229 | 215 | } |
|
230 | 216 | } |
|
231 | 217 | |
|
232 | 218 | lazy_static! { |
|
233 | 219 | static ref SYNTAXES: HashMap<&'static [u8], &'static [u8]> = { |
|
234 | 220 | let mut m = HashMap::new(); |
|
235 | 221 | |
|
236 | 222 | m.insert(b"re".as_ref(), b"relre:".as_ref()); |
|
237 | 223 | m.insert(b"regexp".as_ref(), b"relre:".as_ref()); |
|
238 | 224 | m.insert(b"glob".as_ref(), b"relglob:".as_ref()); |
|
239 | 225 | m.insert(b"rootglob".as_ref(), b"rootglob:".as_ref()); |
|
240 | 226 | m.insert(b"include".as_ref(), b"include".as_ref()); |
|
241 | 227 | m.insert(b"subinclude".as_ref(), b"subinclude".as_ref()); |
|
242 | 228 | m |
|
243 | 229 | }; |
|
244 | 230 | } |
|
245 | 231 | |
|
246 | 232 | pub type PatternTuple = (Vec<u8>, LineNumber, Vec<u8>); |
|
247 | 233 | type WarningTuple = (Vec<u8>, Vec<u8>); |
|
248 | 234 | |
|
249 | 235 | pub fn parse_pattern_file_contents( |
|
250 | 236 | lines: &[u8], |
|
251 | 237 | file_path: &[u8], |
|
252 | 238 | warn: bool, |
|
253 | 239 | ) -> (Vec<PatternTuple>, Vec<WarningTuple>) { |
|
254 | 240 | let comment_regex = Regex::new(r"((?:^|[^\\])(?:\\\\)*)#.*").unwrap(); |
|
255 | 241 | let comment_escape_regex = Regex::new(r"\\#").unwrap(); |
|
256 | 242 | let mut inputs: Vec<PatternTuple> = vec![]; |
|
257 | 243 | let mut warnings: Vec<WarningTuple> = vec![]; |
|
258 | 244 | |
|
259 | 245 | let mut current_syntax = b"relre:".as_ref(); |
|
260 | 246 | |
|
261 | 247 | for (line_number, mut line) in lines.split(|c| *c == b'\n').enumerate() { |
|
262 | 248 | let line_number = line_number + 1; |
|
263 | 249 | |
|
264 | 250 | let line_buf; |
|
265 | 251 | if line.contains(&b'#') { |
|
266 | 252 | if let Some(cap) = comment_regex.captures(line) { |
|
267 | 253 | line = &line[..cap.get(1).unwrap().end()] |
|
268 | 254 | } |
|
269 | 255 | line_buf = comment_escape_regex.replace_all(line, NoExpand(b"#")); |
|
270 | 256 | line = &line_buf; |
|
271 | 257 | } |
|
272 | 258 | |
|
273 | 259 | let mut line = line.trim_end(); |
|
274 | 260 | |
|
275 | 261 | if line.is_empty() { |
|
276 | 262 | continue; |
|
277 | 263 | } |
|
278 | 264 | |
|
279 | 265 | if let Some(syntax) = line.drop_prefix(b"syntax:") { |
|
280 | 266 | let syntax = syntax.trim(); |
|
281 | 267 | |
|
282 | 268 | if let Some(rel_syntax) = SYNTAXES.get(syntax) { |
|
283 | 269 | current_syntax = rel_syntax; |
|
284 | 270 | } else if warn { |
|
285 | 271 | warnings.push((file_path.to_owned(), syntax.to_owned())); |
|
286 | 272 | } |
|
287 | 273 | continue; |
|
288 | 274 | } |
|
289 | 275 | |
|
290 | 276 | let mut line_syntax: &[u8] = ¤t_syntax; |
|
291 | 277 | |
|
292 | 278 | for (s, rels) in SYNTAXES.iter() { |
|
293 | 279 | if let Some(rest) = line.drop_prefix(rels) { |
|
294 | 280 | line_syntax = rels; |
|
295 | 281 | line = rest; |
|
296 | 282 | break; |
|
297 | 283 | } |
|
298 | 284 | if let Some(rest) = line.drop_prefix(&[s, &b":"[..]].concat()) { |
|
299 | 285 | line_syntax = rels; |
|
300 | 286 | line = rest; |
|
301 | 287 | break; |
|
302 | 288 | } |
|
303 | 289 | } |
|
304 | 290 | |
|
305 | 291 | inputs.push(( |
|
306 | 292 | [line_syntax, line].concat(), |
|
307 | 293 | line_number, |
|
308 | 294 | line.to_owned(), |
|
309 | 295 | )); |
|
310 | 296 | } |
|
311 | 297 | (inputs, warnings) |
|
312 | 298 | } |
|
313 | 299 | |
|
314 | 300 | pub fn read_pattern_file( |
|
315 | 301 | file_path: &[u8], |
|
316 | 302 | warn: bool, |
|
317 | 303 | ) -> Result<(Vec<PatternTuple>, Vec<WarningTuple>), PatternFileError> { |
|
318 | 304 | let mut f = File::open(get_path_from_bytes(file_path))?; |
|
319 | 305 | let mut contents = Vec::new(); |
|
320 | 306 | |
|
321 | 307 | f.read_to_end(&mut contents)?; |
|
322 | 308 | |
|
323 | 309 | Ok(parse_pattern_file_contents(&contents, file_path, warn)) |
|
324 | 310 | } |
|
325 | 311 | |
|
326 | 312 | #[cfg(test)] |
|
327 | 313 | mod tests { |
|
328 | 314 | use super::*; |
|
329 | 315 | |
|
330 | 316 | #[test] |
|
331 | 317 | fn escape_pattern_test() { |
|
332 | 318 | let untouched = br#"!"%',/0123456789:;<=>@ABCDEFGHIJKLMNOPQRSTUVWXYZ_`abcdefghijklmnopqrstuvwxyz"#; |
|
333 | 319 | assert_eq!(escape_pattern(untouched), untouched.to_vec()); |
|
334 | 320 | // All escape codes |
|
335 | 321 | assert_eq!( |
|
336 | 322 | escape_pattern(br#"()[]{}?*+-|^$\\.&~# \t\n\r\v\f"#), |
|
337 | 323 | br#"\(\)\[\]\{\}\?\*\+\-\|\^\$\\\\\.\&\~\#\ \\t\\n\\r\\v\\f"# |
|
338 | 324 | .to_vec() |
|
339 | 325 | ); |
|
340 | 326 | } |
|
341 | 327 | |
|
342 | 328 | #[test] |
|
343 | 329 | fn glob_test() { |
|
344 | 330 | assert_eq!(glob_to_re(br#"?"#), br#"."#); |
|
345 | 331 | assert_eq!(glob_to_re(br#"*"#), br#"[^/]*"#); |
|
346 | 332 | assert_eq!(glob_to_re(br#"**"#), br#".*"#); |
|
347 | 333 | assert_eq!(glob_to_re(br#"**/a"#), br#"(?:.*/)?a"#); |
|
348 | 334 | assert_eq!(glob_to_re(br#"a/**/b"#), br#"a/(?:.*/)?b"#); |
|
349 | 335 | assert_eq!(glob_to_re(br#"[a*?!^][^b][!c]"#), br#"[a*?!^][\^b][^c]"#); |
|
350 | 336 | assert_eq!(glob_to_re(br#"{a,b}"#), br#"(?:a|b)"#); |
|
351 | 337 | assert_eq!(glob_to_re(br#".\*\?"#), br#"\.\*\?"#); |
|
352 | 338 | } |
|
353 | 339 | |
|
354 | 340 | #[test] |
|
355 | 341 | fn test_parse_pattern_file_contents() { |
|
356 | 342 | let lines = b"syntax: glob\n*.elc"; |
|
357 | 343 | |
|
358 | 344 | assert_eq!( |
|
359 | 345 | vec![(b"relglob:*.elc".to_vec(), 2, b"*.elc".to_vec())], |
|
360 | 346 | parse_pattern_file_contents(lines, b"file_path", false).0, |
|
361 | 347 | ); |
|
362 | 348 | |
|
363 | 349 | let lines = b"syntax: include\nsyntax: glob"; |
|
364 | 350 | |
|
365 | 351 | assert_eq!( |
|
366 | 352 | parse_pattern_file_contents(lines, b"file_path", false).0, |
|
367 | 353 | vec![] |
|
368 | 354 | ); |
|
369 | 355 | let lines = b"glob:**.o"; |
|
370 | 356 | assert_eq!( |
|
371 | 357 | parse_pattern_file_contents(lines, b"file_path", false).0, |
|
372 | 358 | vec![(b"relglob:**.o".to_vec(), 1, b"**.o".to_vec())] |
|
373 | 359 | ); |
|
374 | 360 | } |
|
375 | 361 | |
|
376 | 362 | #[test] |
|
377 | 363 | fn test_build_single_regex_shortcut() { |
|
378 | 364 | assert_eq!( |
|
379 | 365 | br"(?:/|$)".to_vec(), |
|
380 | 366 | build_single_regex(b"rootglob", b"", b"").unwrap() |
|
381 | 367 | ); |
|
382 | 368 | assert_eq!( |
|
383 | 369 | br"whatever(?:/|$)".to_vec(), |
|
384 | 370 | build_single_regex(b"rootglob", b"whatever", b"").unwrap() |
|
385 | 371 | ); |
|
386 | 372 | assert_eq!( |
|
387 | 373 | br"[^/]*\.o".to_vec(), |
|
388 | 374 | build_single_regex(b"rootglob", b"*.o", b"").unwrap() |
|
389 | 375 | ); |
|
390 | 376 | } |
|
391 | 377 | } |
General Comments 0
You need to be logged in to leave comments.
Login now