Show More
@@ -1,863 +1,863 b'' | |||
|
1 | 1 | use crate::dirstate::entry::TruncatedTimestamp; |
|
2 | 2 | use crate::dirstate::status::IgnoreFnType; |
|
3 | 3 | use crate::dirstate::status::StatusPath; |
|
4 | 4 | use crate::dirstate_tree::dirstate_map::BorrowedPath; |
|
5 | 5 | use crate::dirstate_tree::dirstate_map::ChildNodesRef; |
|
6 | 6 | use crate::dirstate_tree::dirstate_map::DirstateMap; |
|
7 | 7 | use crate::dirstate_tree::dirstate_map::DirstateVersion; |
|
8 | 8 | use crate::dirstate_tree::dirstate_map::NodeRef; |
|
9 | 9 | use crate::dirstate_tree::on_disk::DirstateV2ParseError; |
|
10 | 10 | use crate::matchers::get_ignore_function; |
|
11 | 11 | use crate::matchers::Matcher; |
|
12 | 12 | use crate::utils::files::get_bytes_from_os_string; |
|
13 | 13 | use crate::utils::files::get_path_from_bytes; |
|
14 | 14 | use crate::utils::hg_path::HgPath; |
|
15 | 15 | use crate::BadMatch; |
|
16 | 16 | use crate::DirstateStatus; |
|
17 | 17 | use crate::HgPathBuf; |
|
18 | 18 | use crate::HgPathCow; |
|
19 | 19 | use crate::PatternFileWarning; |
|
20 | 20 | use crate::StatusError; |
|
21 | 21 | use crate::StatusOptions; |
|
22 | 22 | use micro_timer::timed; |
|
23 | 23 | use rayon::prelude::*; |
|
24 | 24 | use sha1::{Digest, Sha1}; |
|
25 | 25 | use std::borrow::Cow; |
|
26 | 26 | use std::io; |
|
27 | 27 | use std::path::Path; |
|
28 | 28 | use std::path::PathBuf; |
|
29 | 29 | use std::sync::Mutex; |
|
30 | 30 | use std::time::SystemTime; |
|
31 | 31 | |
|
32 | 32 | /// Returns the status of the working directory compared to its parent |
|
33 | 33 | /// changeset. |
|
34 | 34 | /// |
|
35 | 35 | /// This algorithm is based on traversing the filesystem tree (`fs` in function |
|
36 | 36 | /// and variable names) and dirstate tree at the same time. The core of this |
|
37 | 37 | /// traversal is the recursive `traverse_fs_directory_and_dirstate` function |
|
38 | 38 | /// and its use of `itertools::merge_join_by`. When reaching a path that only |
|
39 | 39 | /// exists in one of the two trees, depending on information requested by |
|
40 | 40 | /// `options` we may need to traverse the remaining subtree. |
|
41 | 41 | #[timed] |
|
42 | 42 | pub fn status<'dirstate>( |
|
43 | 43 | dmap: &'dirstate mut DirstateMap, |
|
44 | 44 | matcher: &(dyn Matcher + Sync), |
|
45 | 45 | root_dir: PathBuf, |
|
46 | 46 | ignore_files: Vec<PathBuf>, |
|
47 | 47 | options: StatusOptions, |
|
48 | 48 | ) -> Result<(DirstateStatus<'dirstate>, Vec<PatternFileWarning>), StatusError> |
|
49 | 49 | { |
|
50 | 50 | // Force the global rayon threadpool to not exceed 16 concurrent threads. |
|
51 | 51 | // This is a stop-gap measure until we figure out why using more than 16 |
|
52 | 52 | // threads makes `status` slower for each additional thread. |
|
53 | 53 | // We use `ok()` in case the global threadpool has already been |
|
54 | 54 | // instantiated in `rhg` or some other caller. |
|
55 | 55 | // TODO find the underlying cause and fix it, then remove this. |
|
56 | 56 | rayon::ThreadPoolBuilder::new() |
|
57 | 57 | .num_threads(16) |
|
58 | 58 | .build_global() |
|
59 | 59 | .ok(); |
|
60 | 60 | |
|
61 | 61 | let (ignore_fn, warnings, patterns_changed): (IgnoreFnType, _, _) = |
|
62 | 62 | if options.list_ignored || options.list_unknown { |
|
63 | 63 | let (ignore_fn, warnings, changed) = match dmap.dirstate_version { |
|
64 | 64 | DirstateVersion::V1 => { |
|
65 | 65 | let (ignore_fn, warnings) = get_ignore_function( |
|
66 | 66 | ignore_files, |
|
67 | 67 | &root_dir, |
|
68 | 68 | &mut |_pattern_bytes| {}, |
|
69 | 69 | )?; |
|
70 | 70 | (ignore_fn, warnings, None) |
|
71 | 71 | } |
|
72 | 72 | DirstateVersion::V2 => { |
|
73 | 73 | let mut hasher = Sha1::new(); |
|
74 | 74 | let (ignore_fn, warnings) = get_ignore_function( |
|
75 | 75 | ignore_files, |
|
76 | 76 | &root_dir, |
|
77 | 77 | &mut |pattern_bytes| hasher.update(pattern_bytes), |
|
78 | 78 | )?; |
|
79 | 79 | let new_hash = *hasher.finalize().as_ref(); |
|
80 | 80 | let changed = new_hash != dmap.ignore_patterns_hash; |
|
81 | 81 | dmap.ignore_patterns_hash = new_hash; |
|
82 | 82 | (ignore_fn, warnings, Some(changed)) |
|
83 | 83 | } |
|
84 | 84 | }; |
|
85 | 85 | (ignore_fn, warnings, changed) |
|
86 | 86 | } else { |
|
87 | 87 | (Box::new(|&_| true), vec![], None) |
|
88 | 88 | }; |
|
89 | 89 | |
|
90 | 90 | let filesystem_time_at_status_start = |
|
91 | 91 | filesystem_now(&root_dir).ok().map(TruncatedTimestamp::from); |
|
92 | 92 | |
|
93 | 93 | // If the repository is under the current directory, prefer using a |
|
94 | 94 | // relative path, so the kernel needs to traverse fewer directory in every |
|
95 | 95 | // call to `read_dir` or `symlink_metadata`. |
|
96 | 96 | // This is effective in the common case where the current directory is the |
|
97 | 97 | // repository root. |
|
98 | 98 | |
|
99 | 99 | // TODO: Better yet would be to use libc functions like `openat` and |
|
100 | 100 | // `fstatat` to remove such repeated traversals entirely, but the standard |
|
101 | 101 | // library does not provide APIs based on those. |
|
102 | 102 | // Maybe with a crate like https://crates.io/crates/openat instead? |
|
103 | 103 | let root_dir = if let Some(relative) = std::env::current_dir() |
|
104 | 104 | .ok() |
|
105 | 105 | .and_then(|cwd| root_dir.strip_prefix(cwd).ok()) |
|
106 | 106 | { |
|
107 | 107 | relative |
|
108 | 108 | } else { |
|
109 | 109 | &root_dir |
|
110 | 110 | }; |
|
111 | 111 | |
|
112 | 112 | let outcome = DirstateStatus { |
|
113 | 113 | filesystem_time_at_status_start, |
|
114 | 114 | ..Default::default() |
|
115 | 115 | }; |
|
116 | 116 | let common = StatusCommon { |
|
117 | 117 | dmap, |
|
118 | 118 | options, |
|
119 | 119 | matcher, |
|
120 | 120 | ignore_fn, |
|
121 | 121 | outcome: Mutex::new(outcome), |
|
122 | 122 | ignore_patterns_have_changed: patterns_changed, |
|
123 | 123 | new_cachable_directories: Default::default(), |
|
124 | 124 | outated_cached_directories: Default::default(), |
|
125 | 125 | filesystem_time_at_status_start, |
|
126 | 126 | }; |
|
127 | 127 | let is_at_repo_root = true; |
|
128 | 128 | let hg_path = &BorrowedPath::OnDisk(HgPath::new("")); |
|
129 | 129 | let has_ignored_ancestor = false; |
|
130 | 130 | let root_cached_mtime = None; |
|
131 | 131 | let root_dir_metadata = None; |
|
132 | 132 | // If the path we have for the repository root is a symlink, do follow it. |
|
133 | 133 | // (As opposed to symlinks within the working directory which are not |
|
134 | 134 | // followed, using `std::fs::symlink_metadata`.) |
|
135 | 135 | common.traverse_fs_directory_and_dirstate( |
|
136 | 136 | has_ignored_ancestor, |
|
137 | 137 | dmap.root.as_ref(), |
|
138 | 138 | hg_path, |
|
139 | 139 | &root_dir, |
|
140 | 140 | root_dir_metadata, |
|
141 | 141 | root_cached_mtime, |
|
142 | 142 | is_at_repo_root, |
|
143 | 143 | )?; |
|
144 | 144 | let mut outcome = common.outcome.into_inner().unwrap(); |
|
145 | 145 | let new_cachable = common.new_cachable_directories.into_inner().unwrap(); |
|
146 | 146 | let outdated = common.outated_cached_directories.into_inner().unwrap(); |
|
147 | 147 | |
|
148 | 148 | outcome.dirty = common.ignore_patterns_have_changed == Some(true) |
|
149 | 149 | || !outdated.is_empty() |
|
150 | 150 | || (!new_cachable.is_empty() |
|
151 | 151 | && dmap.dirstate_version == DirstateVersion::V2); |
|
152 | 152 | |
|
153 | 153 | // Remove outdated mtimes before adding new mtimes, in case a given |
|
154 | 154 | // directory is both |
|
155 | 155 | for path in &outdated { |
|
156 | 156 | dmap.clear_cached_mtime(path)?; |
|
157 | 157 | } |
|
158 | 158 | for (path, mtime) in &new_cachable { |
|
159 | 159 | dmap.set_cached_mtime(path, *mtime)?; |
|
160 | 160 | } |
|
161 | 161 | |
|
162 | 162 | Ok((outcome, warnings)) |
|
163 | 163 | } |
|
164 | 164 | |
|
165 | 165 | /// Bag of random things needed by various parts of the algorithm. Reduces the |
|
166 | 166 | /// number of parameters passed to functions. |
|
167 | 167 | struct StatusCommon<'a, 'tree, 'on_disk: 'tree> { |
|
168 | 168 | dmap: &'tree DirstateMap<'on_disk>, |
|
169 | 169 | options: StatusOptions, |
|
170 | 170 | matcher: &'a (dyn Matcher + Sync), |
|
171 | 171 | ignore_fn: IgnoreFnType<'a>, |
|
172 | 172 | outcome: Mutex<DirstateStatus<'on_disk>>, |
|
173 | 173 | new_cachable_directories: |
|
174 | 174 | Mutex<Vec<(Cow<'on_disk, HgPath>, TruncatedTimestamp)>>, |
|
175 | 175 | outated_cached_directories: Mutex<Vec<Cow<'on_disk, HgPath>>>, |
|
176 | 176 | |
|
177 | 177 | /// Whether ignore files like `.hgignore` have changed since the previous |
|
178 | 178 | /// time a `status()` call wrote their hash to the dirstate. `None` means |
|
179 | 179 | /// we don’t know as this run doesn’t list either ignored or uknown files |
|
180 | 180 | /// and therefore isn’t reading `.hgignore`. |
|
181 | 181 | ignore_patterns_have_changed: Option<bool>, |
|
182 | 182 | |
|
183 | 183 | /// The current time at the start of the `status()` algorithm, as measured |
|
184 | 184 | /// and possibly truncated by the filesystem. |
|
185 | 185 | filesystem_time_at_status_start: Option<TruncatedTimestamp>, |
|
186 | 186 | } |
|
187 | 187 | |
|
188 | 188 | enum Outcome { |
|
189 | 189 | Modified, |
|
190 | 190 | Added, |
|
191 | 191 | Removed, |
|
192 | 192 | Deleted, |
|
193 | 193 | Clean, |
|
194 | 194 | Ignored, |
|
195 | 195 | Unknown, |
|
196 | 196 | Unsure, |
|
197 | 197 | } |
|
198 | 198 | |
|
199 | 199 | impl<'a, 'tree, 'on_disk> StatusCommon<'a, 'tree, 'on_disk> { |
|
200 | 200 | fn push_outcome( |
|
201 | 201 | &self, |
|
202 | 202 | which: Outcome, |
|
203 | 203 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
204 | 204 | ) -> Result<(), DirstateV2ParseError> { |
|
205 | 205 | let path = dirstate_node |
|
206 | 206 | .full_path_borrowed(self.dmap.on_disk)? |
|
207 | 207 | .detach_from_tree(); |
|
208 | 208 | let copy_source = if self.options.list_copies { |
|
209 | 209 | dirstate_node |
|
210 | 210 | .copy_source_borrowed(self.dmap.on_disk)? |
|
211 | 211 | .map(|source| source.detach_from_tree()) |
|
212 | 212 | } else { |
|
213 | 213 | None |
|
214 | 214 | }; |
|
215 | 215 | self.push_outcome_common(which, path, copy_source); |
|
216 | 216 | Ok(()) |
|
217 | 217 | } |
|
218 | 218 | |
|
219 | 219 | fn push_outcome_without_copy_source( |
|
220 | 220 | &self, |
|
221 | 221 | which: Outcome, |
|
222 | 222 | path: &BorrowedPath<'_, 'on_disk>, |
|
223 | 223 | ) { |
|
224 | 224 | self.push_outcome_common(which, path.detach_from_tree(), None) |
|
225 | 225 | } |
|
226 | 226 | |
|
227 | 227 | fn push_outcome_common( |
|
228 | 228 | &self, |
|
229 | 229 | which: Outcome, |
|
230 | 230 | path: HgPathCow<'on_disk>, |
|
231 | 231 | copy_source: Option<HgPathCow<'on_disk>>, |
|
232 | 232 | ) { |
|
233 | 233 | let mut outcome = self.outcome.lock().unwrap(); |
|
234 | 234 | let vec = match which { |
|
235 | 235 | Outcome::Modified => &mut outcome.modified, |
|
236 | 236 | Outcome::Added => &mut outcome.added, |
|
237 | 237 | Outcome::Removed => &mut outcome.removed, |
|
238 | 238 | Outcome::Deleted => &mut outcome.deleted, |
|
239 | 239 | Outcome::Clean => &mut outcome.clean, |
|
240 | 240 | Outcome::Ignored => &mut outcome.ignored, |
|
241 | 241 | Outcome::Unknown => &mut outcome.unknown, |
|
242 | 242 | Outcome::Unsure => &mut outcome.unsure, |
|
243 | 243 | }; |
|
244 | 244 | vec.push(StatusPath { path, copy_source }); |
|
245 | 245 | } |
|
246 | 246 | |
|
247 | 247 | fn read_dir( |
|
248 | 248 | &self, |
|
249 | 249 | hg_path: &HgPath, |
|
250 | 250 | fs_path: &Path, |
|
251 | 251 | is_at_repo_root: bool, |
|
252 | 252 | ) -> Result<Vec<DirEntry>, ()> { |
|
253 | 253 | DirEntry::read_dir(fs_path, is_at_repo_root) |
|
254 | 254 | .map_err(|error| self.io_error(error, hg_path)) |
|
255 | 255 | } |
|
256 | 256 | |
|
257 | 257 | fn io_error(&self, error: std::io::Error, hg_path: &HgPath) { |
|
258 | 258 | let errno = error.raw_os_error().expect("expected real OS error"); |
|
259 | 259 | self.outcome |
|
260 | 260 | .lock() |
|
261 | 261 | .unwrap() |
|
262 | 262 | .bad |
|
263 | 263 | .push((hg_path.to_owned().into(), BadMatch::OsError(errno))) |
|
264 | 264 | } |
|
265 | 265 | |
|
266 | 266 | fn check_for_outdated_directory_cache( |
|
267 | 267 | &self, |
|
268 | 268 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
269 | 269 | ) -> Result<(), DirstateV2ParseError> { |
|
270 | 270 | if self.ignore_patterns_have_changed == Some(true) |
|
271 | 271 | && dirstate_node.cached_directory_mtime()?.is_some() |
|
272 | 272 | { |
|
273 | 273 | self.outated_cached_directories.lock().unwrap().push( |
|
274 | 274 | dirstate_node |
|
275 | 275 | .full_path_borrowed(self.dmap.on_disk)? |
|
276 | 276 | .detach_from_tree(), |
|
277 | 277 | ) |
|
278 | 278 | } |
|
279 | 279 | Ok(()) |
|
280 | 280 | } |
|
281 | 281 | |
|
282 | 282 | /// If this returns true, we can get accurate results by only using |
|
283 | 283 | /// `symlink_metadata` for child nodes that exist in the dirstate and don’t |
|
284 | 284 | /// need to call `read_dir`. |
|
285 | 285 | fn can_skip_fs_readdir( |
|
286 | 286 | &self, |
|
287 | 287 | directory_metadata: Option<&std::fs::Metadata>, |
|
288 | 288 | cached_directory_mtime: Option<TruncatedTimestamp>, |
|
289 | 289 | ) -> bool { |
|
290 | 290 | if !self.options.list_unknown && !self.options.list_ignored { |
|
291 | 291 | // All states that we care about listing have corresponding |
|
292 | 292 | // dirstate entries. |
|
293 | 293 | // This happens for example with `hg status -mard`. |
|
294 | 294 | return true; |
|
295 | 295 | } |
|
296 | 296 | if !self.options.list_ignored |
|
297 | 297 | && self.ignore_patterns_have_changed == Some(false) |
|
298 | 298 | { |
|
299 | 299 | if let Some(cached_mtime) = cached_directory_mtime { |
|
300 | 300 | // The dirstate contains a cached mtime for this directory, set |
|
301 | 301 | // by a previous run of the `status` algorithm which found this |
|
302 | 302 | // directory eligible for `read_dir` caching. |
|
303 | 303 | if let Some(meta) = directory_metadata { |
|
304 | 304 | if cached_mtime |
|
305 | 305 | .likely_equal_to_mtime_of(meta) |
|
306 | 306 | .unwrap_or(false) |
|
307 | 307 | { |
|
308 | 308 | // The mtime of that directory has not changed |
|
309 | 309 | // since then, which means that the results of |
|
310 | 310 | // `read_dir` should also be unchanged. |
|
311 | 311 | return true; |
|
312 | 312 | } |
|
313 | 313 | } |
|
314 | 314 | } |
|
315 | 315 | } |
|
316 | 316 | false |
|
317 | 317 | } |
|
318 | 318 | |
|
319 | 319 | /// Returns whether all child entries of the filesystem directory have a |
|
320 | 320 | /// corresponding dirstate node or are ignored. |
|
321 | 321 | fn traverse_fs_directory_and_dirstate( |
|
322 | 322 | &self, |
|
323 | 323 | has_ignored_ancestor: bool, |
|
324 | 324 | dirstate_nodes: ChildNodesRef<'tree, 'on_disk>, |
|
325 | 325 | directory_hg_path: &BorrowedPath<'tree, 'on_disk>, |
|
326 | 326 | directory_fs_path: &Path, |
|
327 | 327 | directory_metadata: Option<&std::fs::Metadata>, |
|
328 | 328 | cached_directory_mtime: Option<TruncatedTimestamp>, |
|
329 | 329 | is_at_repo_root: bool, |
|
330 | 330 | ) -> Result<bool, DirstateV2ParseError> { |
|
331 | 331 | if self.can_skip_fs_readdir(directory_metadata, cached_directory_mtime) |
|
332 | 332 | { |
|
333 | 333 | dirstate_nodes |
|
334 | 334 | .par_iter() |
|
335 | 335 | .map(|dirstate_node| { |
|
336 | 336 | let fs_path = directory_fs_path.join(get_path_from_bytes( |
|
337 | 337 | dirstate_node.base_name(self.dmap.on_disk)?.as_bytes(), |
|
338 | 338 | )); |
|
339 | 339 | match std::fs::symlink_metadata(&fs_path) { |
|
340 | 340 | Ok(fs_metadata) => self.traverse_fs_and_dirstate( |
|
341 | 341 | &fs_path, |
|
342 | 342 | &fs_metadata, |
|
343 | 343 | dirstate_node, |
|
344 | 344 | has_ignored_ancestor, |
|
345 | 345 | ), |
|
346 | 346 | Err(e) if e.kind() == std::io::ErrorKind::NotFound => { |
|
347 | 347 | self.traverse_dirstate_only(dirstate_node) |
|
348 | 348 | } |
|
349 | 349 | Err(error) => { |
|
350 | 350 | let hg_path = |
|
351 | 351 | dirstate_node.full_path(self.dmap.on_disk)?; |
|
352 | 352 | Ok(self.io_error(error, hg_path)) |
|
353 | 353 | } |
|
354 | 354 | } |
|
355 | 355 | }) |
|
356 | 356 | .collect::<Result<_, _>>()?; |
|
357 | 357 | |
|
358 | 358 | // We don’t know, so conservatively say this isn’t the case |
|
359 | 359 | let children_all_have_dirstate_node_or_are_ignored = false; |
|
360 | 360 | |
|
361 | 361 | return Ok(children_all_have_dirstate_node_or_are_ignored); |
|
362 | 362 | } |
|
363 | 363 | |
|
364 | 364 | let mut fs_entries = if let Ok(entries) = self.read_dir( |
|
365 | 365 | directory_hg_path, |
|
366 | 366 | directory_fs_path, |
|
367 | 367 | is_at_repo_root, |
|
368 | 368 | ) { |
|
369 | 369 | entries |
|
370 | 370 | } else { |
|
371 | 371 | // Treat an unreadable directory (typically because of insufficient |
|
372 | 372 | // permissions) like an empty directory. `self.read_dir` has |
|
373 | 373 | // already called `self.io_error` so a warning will be emitted. |
|
374 | 374 | Vec::new() |
|
375 | 375 | }; |
|
376 | 376 | |
|
377 | 377 | // `merge_join_by` requires both its input iterators to be sorted: |
|
378 | 378 | |
|
379 | 379 | let dirstate_nodes = dirstate_nodes.sorted(); |
|
380 | 380 | // `sort_unstable_by_key` doesn’t allow keys borrowing from the value: |
|
381 | 381 | // https://github.com/rust-lang/rust/issues/34162 |
|
382 | 382 | fs_entries.sort_unstable_by(|e1, e2| e1.base_name.cmp(&e2.base_name)); |
|
383 | 383 | |
|
384 | 384 | // Propagate here any error that would happen inside the comparison |
|
385 | 385 | // callback below |
|
386 | 386 | for dirstate_node in &dirstate_nodes { |
|
387 | 387 | dirstate_node.base_name(self.dmap.on_disk)?; |
|
388 | 388 | } |
|
389 | 389 | itertools::merge_join_by( |
|
390 | 390 | dirstate_nodes, |
|
391 | 391 | &fs_entries, |
|
392 | 392 | |dirstate_node, fs_entry| { |
|
393 | 393 | // This `unwrap` never panics because we already propagated |
|
394 | 394 | // those errors above |
|
395 | 395 | dirstate_node |
|
396 | 396 | .base_name(self.dmap.on_disk) |
|
397 | 397 | .unwrap() |
|
398 | 398 | .cmp(&fs_entry.base_name) |
|
399 | 399 | }, |
|
400 | 400 | ) |
|
401 | 401 | .par_bridge() |
|
402 | 402 | .map(|pair| { |
|
403 | 403 | use itertools::EitherOrBoth::*; |
|
404 | 404 | let has_dirstate_node_or_is_ignored; |
|
405 | 405 | match pair { |
|
406 | 406 | Both(dirstate_node, fs_entry) => { |
|
407 | 407 | self.traverse_fs_and_dirstate( |
|
408 | 408 | &fs_entry.full_path, |
|
409 | 409 | &fs_entry.metadata, |
|
410 | 410 | dirstate_node, |
|
411 | 411 | has_ignored_ancestor, |
|
412 | 412 | )?; |
|
413 | 413 | has_dirstate_node_or_is_ignored = true |
|
414 | 414 | } |
|
415 | 415 | Left(dirstate_node) => { |
|
416 | 416 | self.traverse_dirstate_only(dirstate_node)?; |
|
417 | 417 | has_dirstate_node_or_is_ignored = true; |
|
418 | 418 | } |
|
419 | 419 | Right(fs_entry) => { |
|
420 | 420 | has_dirstate_node_or_is_ignored = self.traverse_fs_only( |
|
421 | 421 | has_ignored_ancestor, |
|
422 | 422 | directory_hg_path, |
|
423 | 423 | fs_entry, |
|
424 | 424 | ) |
|
425 | 425 | } |
|
426 | 426 | } |
|
427 | 427 | Ok(has_dirstate_node_or_is_ignored) |
|
428 | 428 | }) |
|
429 | 429 | .try_reduce(|| true, |a, b| Ok(a && b)) |
|
430 | 430 | } |
|
431 | 431 | |
|
432 | 432 | fn traverse_fs_and_dirstate( |
|
433 | 433 | &self, |
|
434 | 434 | fs_path: &Path, |
|
435 | 435 | fs_metadata: &std::fs::Metadata, |
|
436 | 436 | dirstate_node: NodeRef<'tree, 'on_disk>, |
|
437 | 437 | has_ignored_ancestor: bool, |
|
438 | 438 | ) -> Result<(), DirstateV2ParseError> { |
|
439 | 439 | self.check_for_outdated_directory_cache(&dirstate_node)?; |
|
440 | 440 | let hg_path = &dirstate_node.full_path_borrowed(self.dmap.on_disk)?; |
|
441 | 441 | let file_type = fs_metadata.file_type(); |
|
442 | 442 | let file_or_symlink = file_type.is_file() || file_type.is_symlink(); |
|
443 | 443 | if !file_or_symlink { |
|
444 | 444 | // If we previously had a file here, it was removed (with |
|
445 | 445 | // `hg rm` or similar) or deleted before it could be |
|
446 | 446 | // replaced by a directory or something else. |
|
447 | 447 | self.mark_removed_or_deleted_if_file(&dirstate_node)?; |
|
448 | 448 | } |
|
449 | 449 | if file_type.is_dir() { |
|
450 | 450 | if self.options.collect_traversed_dirs { |
|
451 | 451 | self.outcome |
|
452 | 452 | .lock() |
|
453 | 453 | .unwrap() |
|
454 | 454 | .traversed |
|
455 | 455 | .push(hg_path.detach_from_tree()) |
|
456 | 456 | } |
|
457 | 457 | let is_ignored = has_ignored_ancestor || (self.ignore_fn)(hg_path); |
|
458 | 458 | let is_at_repo_root = false; |
|
459 | 459 | let children_all_have_dirstate_node_or_are_ignored = self |
|
460 | 460 | .traverse_fs_directory_and_dirstate( |
|
461 | 461 | is_ignored, |
|
462 | 462 | dirstate_node.children(self.dmap.on_disk)?, |
|
463 | 463 | hg_path, |
|
464 | 464 | fs_path, |
|
465 | 465 | Some(fs_metadata), |
|
466 | 466 | dirstate_node.cached_directory_mtime()?, |
|
467 | 467 | is_at_repo_root, |
|
468 | 468 | )?; |
|
469 | 469 | self.maybe_save_directory_mtime( |
|
470 | 470 | children_all_have_dirstate_node_or_are_ignored, |
|
471 | 471 | fs_metadata, |
|
472 | 472 | dirstate_node, |
|
473 | 473 | )? |
|
474 | 474 | } else { |
|
475 | 475 | if file_or_symlink && self.matcher.matches(hg_path) { |
|
476 | 476 | if let Some(entry) = dirstate_node.entry()? { |
|
477 | 477 | if !entry.any_tracked() { |
|
478 | 478 | // Forward-compat if we start tracking unknown/ignored |
|
479 | 479 | // files for caching reasons |
|
480 | 480 | self.mark_unknown_or_ignored( |
|
481 | 481 | has_ignored_ancestor, |
|
482 | 482 | hg_path, |
|
483 | 483 | ); |
|
484 | 484 | } |
|
485 | 485 | if entry.added() { |
|
486 | 486 | self.push_outcome(Outcome::Added, &dirstate_node)?; |
|
487 | 487 | } else if entry.removed() { |
|
488 | 488 | self.push_outcome(Outcome::Removed, &dirstate_node)?; |
|
489 | 489 | } else if entry.modified() { |
|
490 | 490 | self.push_outcome(Outcome::Modified, &dirstate_node)?; |
|
491 | 491 | } else { |
|
492 | 492 | self.handle_normal_file(&dirstate_node, fs_metadata)?; |
|
493 | 493 | } |
|
494 | 494 | } else { |
|
495 | 495 | // `node.entry.is_none()` indicates a "directory" |
|
496 | 496 | // node, but the filesystem has a file |
|
497 | 497 | self.mark_unknown_or_ignored( |
|
498 | 498 | has_ignored_ancestor, |
|
499 | 499 | hg_path, |
|
500 | 500 | ); |
|
501 | 501 | } |
|
502 | 502 | } |
|
503 | 503 | |
|
504 | 504 | for child_node in dirstate_node.children(self.dmap.on_disk)?.iter() |
|
505 | 505 | { |
|
506 | 506 | self.traverse_dirstate_only(child_node)? |
|
507 | 507 | } |
|
508 | 508 | } |
|
509 | 509 | Ok(()) |
|
510 | 510 | } |
|
511 | 511 | |
|
512 | 512 | fn maybe_save_directory_mtime( |
|
513 | 513 | &self, |
|
514 | 514 | children_all_have_dirstate_node_or_are_ignored: bool, |
|
515 | 515 | directory_metadata: &std::fs::Metadata, |
|
516 | 516 | dirstate_node: NodeRef<'tree, 'on_disk>, |
|
517 | 517 | ) -> Result<(), DirstateV2ParseError> { |
|
518 | 518 | if !children_all_have_dirstate_node_or_are_ignored { |
|
519 | 519 | return Ok(()); |
|
520 | 520 | } |
|
521 | 521 | // All filesystem directory entries from `read_dir` have a |
|
522 | 522 | // corresponding node in the dirstate, so we can reconstitute the |
|
523 | 523 | // names of those entries without calling `read_dir` again. |
|
524 | 524 | |
|
525 | 525 | // TODO: use let-else here and below when available: |
|
526 | 526 | // https://github.com/rust-lang/rust/issues/87335 |
|
527 | 527 | let status_start = if let Some(status_start) = |
|
528 | 528 | &self.filesystem_time_at_status_start |
|
529 | 529 | { |
|
530 | 530 | status_start |
|
531 | 531 | } else { |
|
532 | 532 | return Ok(()); |
|
533 | 533 | }; |
|
534 | 534 | |
|
535 | 535 | // Although the Rust standard library’s `SystemTime` type |
|
536 | 536 | // has nanosecond precision, the times reported for a |
|
537 | 537 | // directory’s (or file’s) modified time may have lower |
|
538 | 538 | // resolution based on the filesystem (for example ext3 |
|
539 | 539 | // only stores integer seconds), kernel (see |
|
540 | 540 | // https://stackoverflow.com/a/14393315/1162888), etc. |
|
541 | 541 | let directory_mtime = if let Ok(option) = |
|
542 | 542 | TruncatedTimestamp::for_reliable_mtime_of( |
|
543 | 543 | directory_metadata, |
|
544 | 544 | status_start, |
|
545 | 545 | ) { |
|
546 | 546 | if let Some(directory_mtime) = option { |
|
547 | 547 | directory_mtime |
|
548 | 548 | } else { |
|
549 | 549 | // The directory was modified too recently, |
|
550 | 550 | // don’t cache its `read_dir` results. |
|
551 | 551 | // |
|
552 | 552 | // 1. A change to this directory (direct child was |
|
553 | 553 | // added or removed) cause its mtime to be set |
|
554 | 554 | // (possibly truncated) to `directory_mtime` |
|
555 | 555 | // 2. This `status` algorithm calls `read_dir` |
|
556 | 556 | // 3. An other change is made to the same directory is |
|
557 | 557 | // made so that calling `read_dir` agin would give |
|
558 | 558 | // different results, but soon enough after 1. that |
|
559 | 559 | // the mtime stays the same |
|
560 | 560 | // |
|
561 | 561 | // On a system where the time resolution poor, this |
|
562 | 562 | // scenario is not unlikely if all three steps are caused |
|
563 | 563 | // by the same script. |
|
564 | 564 | return Ok(()); |
|
565 | 565 | } |
|
566 | 566 | } else { |
|
567 | 567 | // OS/libc does not support mtime? |
|
568 | 568 | return Ok(()); |
|
569 | 569 | }; |
|
570 | 570 | // We’ve observed (through `status_start`) that time has |
|
571 | 571 | // “progressed” since `directory_mtime`, so any further |
|
572 | 572 | // change to this directory is extremely likely to cause a |
|
573 | 573 | // different mtime. |
|
574 | 574 | // |
|
575 | 575 | // Having the same mtime again is not entirely impossible |
|
576 | 576 | // since the system clock is not monotonous. It could jump |
|
577 | 577 | // backward to some point before `directory_mtime`, then a |
|
578 | 578 | // directory change could potentially happen during exactly |
|
579 | 579 | // the wrong tick. |
|
580 | 580 | // |
|
581 | 581 | // We deem this scenario (unlike the previous one) to be |
|
582 | 582 | // unlikely enough in practice. |
|
583 | 583 | |
|
584 | 584 | let is_up_to_date = |
|
585 | 585 | if let Some(cached) = dirstate_node.cached_directory_mtime()? { |
|
586 | 586 | cached.likely_equal(directory_mtime) |
|
587 | 587 | } else { |
|
588 | 588 | false |
|
589 | 589 | }; |
|
590 | 590 | if !is_up_to_date { |
|
591 | 591 | let hg_path = dirstate_node |
|
592 | 592 | .full_path_borrowed(self.dmap.on_disk)? |
|
593 | 593 | .detach_from_tree(); |
|
594 | 594 | self.new_cachable_directories |
|
595 | 595 | .lock() |
|
596 | 596 | .unwrap() |
|
597 | 597 | .push((hg_path, directory_mtime)) |
|
598 | 598 | } |
|
599 | 599 | Ok(()) |
|
600 | 600 | } |
|
601 | 601 | |
|
602 | 602 | /// A file that is clean in the dirstate was found in the filesystem |
|
603 | 603 | fn handle_normal_file( |
|
604 | 604 | &self, |
|
605 | 605 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
606 | 606 | fs_metadata: &std::fs::Metadata, |
|
607 | 607 | ) -> Result<(), DirstateV2ParseError> { |
|
608 | 608 | // Keep the low 31 bits |
|
609 | 609 | fn truncate_u64(value: u64) -> i32 { |
|
610 | 610 | (value & 0x7FFF_FFFF) as i32 |
|
611 | 611 | } |
|
612 | 612 | |
|
613 | 613 | let entry = dirstate_node |
|
614 | 614 | .entry()? |
|
615 | 615 | .expect("handle_normal_file called with entry-less node"); |
|
616 | 616 | let mode_changed = |
|
617 | 617 | || self.options.check_exec && entry.mode_changed(fs_metadata); |
|
618 | 618 | let size = entry.size(); |
|
619 | 619 | let size_changed = size != truncate_u64(fs_metadata.len()); |
|
620 | 620 | if size >= 0 && size_changed && fs_metadata.file_type().is_symlink() { |
|
621 | 621 | // issue6456: Size returned may be longer due to encryption |
|
622 | 622 | // on EXT-4 fscrypt. TODO maybe only do it on EXT4? |
|
623 | 623 | self.push_outcome(Outcome::Unsure, dirstate_node)? |
|
624 | 624 | } else if dirstate_node.has_copy_source() |
|
625 | 625 | || entry.is_from_other_parent() |
|
626 | 626 | || (size >= 0 && (size_changed || mode_changed())) |
|
627 | 627 | { |
|
628 | 628 | self.push_outcome(Outcome::Modified, dirstate_node)? |
|
629 | 629 | } else { |
|
630 | 630 | let mtime_looks_clean; |
|
631 | 631 | if let Some(dirstate_mtime) = entry.truncated_mtime() { |
|
632 | 632 | let fs_mtime = TruncatedTimestamp::for_mtime_of(fs_metadata) |
|
633 | 633 | .expect("OS/libc does not support mtime?"); |
|
634 | 634 | // There might be a change in the future if for example the |
|
635 | 635 | // internal clock become off while process run, but this is a |
|
636 | 636 | // case where the issues the user would face |
|
637 | 637 | // would be a lot worse and there is nothing we |
|
638 | 638 | // can really do. |
|
639 | 639 | mtime_looks_clean = fs_mtime.likely_equal(dirstate_mtime) |
|
640 | 640 | } else { |
|
641 | 641 | // No mtime in the dirstate entry |
|
642 | 642 | mtime_looks_clean = false |
|
643 | 643 | }; |
|
644 | 644 | if !mtime_looks_clean { |
|
645 | 645 | self.push_outcome(Outcome::Unsure, dirstate_node)? |
|
646 | 646 | } else if self.options.list_clean { |
|
647 | 647 | self.push_outcome(Outcome::Clean, dirstate_node)? |
|
648 | 648 | } |
|
649 | 649 | } |
|
650 | 650 | Ok(()) |
|
651 | 651 | } |
|
652 | 652 | |
|
653 | 653 | /// A node in the dirstate tree has no corresponding filesystem entry |
|
654 | 654 | fn traverse_dirstate_only( |
|
655 | 655 | &self, |
|
656 | 656 | dirstate_node: NodeRef<'tree, 'on_disk>, |
|
657 | 657 | ) -> Result<(), DirstateV2ParseError> { |
|
658 | 658 | self.check_for_outdated_directory_cache(&dirstate_node)?; |
|
659 | 659 | self.mark_removed_or_deleted_if_file(&dirstate_node)?; |
|
660 | 660 | dirstate_node |
|
661 | 661 | .children(self.dmap.on_disk)? |
|
662 | 662 | .par_iter() |
|
663 | 663 | .map(|child_node| self.traverse_dirstate_only(child_node)) |
|
664 | 664 | .collect() |
|
665 | 665 | } |
|
666 | 666 | |
|
667 | 667 | /// A node in the dirstate tree has no corresponding *file* on the |
|
668 | 668 | /// filesystem |
|
669 | 669 | /// |
|
670 | 670 | /// Does nothing on a "directory" node |
|
671 | 671 | fn mark_removed_or_deleted_if_file( |
|
672 | 672 | &self, |
|
673 | 673 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
674 | 674 | ) -> Result<(), DirstateV2ParseError> { |
|
675 | 675 | if let Some(entry) = dirstate_node.entry()? { |
|
676 | 676 | if !entry.any_tracked() { |
|
677 | 677 | // Future-compat for when we start storing ignored and unknown |
|
678 | 678 | // files for caching reasons |
|
679 | 679 | return Ok(()); |
|
680 | 680 | } |
|
681 | 681 | let path = dirstate_node.full_path(self.dmap.on_disk)?; |
|
682 | 682 | if self.matcher.matches(path) { |
|
683 | 683 | if entry.removed() { |
|
684 | 684 | self.push_outcome(Outcome::Removed, dirstate_node)? |
|
685 | 685 | } else { |
|
686 | 686 | self.push_outcome(Outcome::Deleted, &dirstate_node)? |
|
687 | 687 | } |
|
688 | 688 | } |
|
689 | 689 | } |
|
690 | 690 | Ok(()) |
|
691 | 691 | } |
|
692 | 692 | |
|
693 | 693 | /// Something in the filesystem has no corresponding dirstate node |
|
694 | 694 | /// |
|
695 | 695 | /// Returns whether that path is ignored |
|
696 | 696 | fn traverse_fs_only( |
|
697 | 697 | &self, |
|
698 | 698 | has_ignored_ancestor: bool, |
|
699 | 699 | directory_hg_path: &HgPath, |
|
700 | 700 | fs_entry: &DirEntry, |
|
701 | 701 | ) -> bool { |
|
702 | 702 | let hg_path = directory_hg_path.join(&fs_entry.base_name); |
|
703 | 703 | let file_type = fs_entry.metadata.file_type(); |
|
704 | 704 | let file_or_symlink = file_type.is_file() || file_type.is_symlink(); |
|
705 | 705 | if file_type.is_dir() { |
|
706 | 706 | let is_ignored = |
|
707 | 707 | has_ignored_ancestor || (self.ignore_fn)(&hg_path); |
|
708 | 708 | let traverse_children = if is_ignored { |
|
709 | 709 | // Descendants of an ignored directory are all ignored |
|
710 | 710 | self.options.list_ignored |
|
711 | 711 | } else { |
|
712 | 712 | // Descendants of an unknown directory may be either unknown or |
|
713 | 713 | // ignored |
|
714 | 714 | self.options.list_unknown || self.options.list_ignored |
|
715 | 715 | }; |
|
716 | 716 | if traverse_children { |
|
717 | 717 | let is_at_repo_root = false; |
|
718 | 718 | if let Ok(children_fs_entries) = self.read_dir( |
|
719 | 719 | &hg_path, |
|
720 | 720 | &fs_entry.full_path, |
|
721 | 721 | is_at_repo_root, |
|
722 | 722 | ) { |
|
723 | 723 | children_fs_entries.par_iter().for_each(|child_fs_entry| { |
|
724 | 724 | self.traverse_fs_only( |
|
725 | 725 | is_ignored, |
|
726 | 726 | &hg_path, |
|
727 | 727 | child_fs_entry, |
|
728 | 728 | ); |
|
729 | 729 | }) |
|
730 | 730 | } |
|
731 | } | |
|
732 | if self.options.collect_traversed_dirs { | |
|
733 | self.outcome.lock().unwrap().traversed.push(hg_path.into()) | |
|
731 | if self.options.collect_traversed_dirs { | |
|
732 | self.outcome.lock().unwrap().traversed.push(hg_path.into()) | |
|
733 | } | |
|
734 | 734 | } |
|
735 | 735 | is_ignored |
|
736 | 736 | } else { |
|
737 | 737 | if file_or_symlink { |
|
738 | 738 | if self.matcher.matches(&hg_path) { |
|
739 | 739 | self.mark_unknown_or_ignored( |
|
740 | 740 | has_ignored_ancestor, |
|
741 | 741 | &BorrowedPath::InMemory(&hg_path), |
|
742 | 742 | ) |
|
743 | 743 | } else { |
|
744 | 744 | // We haven’t computed whether this path is ignored. It |
|
745 | 745 | // might not be, and a future run of status might have a |
|
746 | 746 | // different matcher that matches it. So treat it as not |
|
747 | 747 | // ignored. That is, inhibit readdir caching of the parent |
|
748 | 748 | // directory. |
|
749 | 749 | false |
|
750 | 750 | } |
|
751 | 751 | } else { |
|
752 | 752 | // This is neither a directory, a plain file, or a symlink. |
|
753 | 753 | // Treat it like an ignored file. |
|
754 | 754 | true |
|
755 | 755 | } |
|
756 | 756 | } |
|
757 | 757 | } |
|
758 | 758 | |
|
759 | 759 | /// Returns whether that path is ignored |
|
760 | 760 | fn mark_unknown_or_ignored( |
|
761 | 761 | &self, |
|
762 | 762 | has_ignored_ancestor: bool, |
|
763 | 763 | hg_path: &BorrowedPath<'_, 'on_disk>, |
|
764 | 764 | ) -> bool { |
|
765 | 765 | let is_ignored = has_ignored_ancestor || (self.ignore_fn)(&hg_path); |
|
766 | 766 | if is_ignored { |
|
767 | 767 | if self.options.list_ignored { |
|
768 | 768 | self.push_outcome_without_copy_source( |
|
769 | 769 | Outcome::Ignored, |
|
770 | 770 | hg_path, |
|
771 | 771 | ) |
|
772 | 772 | } |
|
773 | 773 | } else { |
|
774 | 774 | if self.options.list_unknown { |
|
775 | 775 | self.push_outcome_without_copy_source( |
|
776 | 776 | Outcome::Unknown, |
|
777 | 777 | hg_path, |
|
778 | 778 | ) |
|
779 | 779 | } |
|
780 | 780 | } |
|
781 | 781 | is_ignored |
|
782 | 782 | } |
|
783 | 783 | } |
|
784 | 784 | |
|
785 | 785 | struct DirEntry { |
|
786 | 786 | base_name: HgPathBuf, |
|
787 | 787 | full_path: PathBuf, |
|
788 | 788 | metadata: std::fs::Metadata, |
|
789 | 789 | } |
|
790 | 790 | |
|
791 | 791 | impl DirEntry { |
|
792 | 792 | /// Returns **unsorted** entries in the given directory, with name and |
|
793 | 793 | /// metadata. |
|
794 | 794 | /// |
|
795 | 795 | /// If a `.hg` sub-directory is encountered: |
|
796 | 796 | /// |
|
797 | 797 | /// * At the repository root, ignore that sub-directory |
|
798 | 798 | /// * Elsewhere, we’re listing the content of a sub-repo. Return an empty |
|
799 | 799 | /// list instead. |
|
800 | 800 | fn read_dir(path: &Path, is_at_repo_root: bool) -> io::Result<Vec<Self>> { |
|
801 | 801 | // `read_dir` returns a "not found" error for the empty path |
|
802 | 802 | let at_cwd = path == Path::new(""); |
|
803 | 803 | let read_dir_path = if at_cwd { Path::new(".") } else { path }; |
|
804 | 804 | let mut results = Vec::new(); |
|
805 | 805 | for entry in read_dir_path.read_dir()? { |
|
806 | 806 | let entry = entry?; |
|
807 | 807 | let metadata = match entry.metadata() { |
|
808 | 808 | Ok(v) => v, |
|
809 | 809 | Err(e) => { |
|
810 | 810 | // race with file deletion? |
|
811 | 811 | if e.kind() == std::io::ErrorKind::NotFound { |
|
812 | 812 | continue; |
|
813 | 813 | } else { |
|
814 | 814 | return Err(e); |
|
815 | 815 | } |
|
816 | 816 | } |
|
817 | 817 | }; |
|
818 | 818 | let file_name = entry.file_name(); |
|
819 | 819 | // FIXME don't do this when cached |
|
820 | 820 | if file_name == ".hg" { |
|
821 | 821 | if is_at_repo_root { |
|
822 | 822 | // Skip the repo’s own .hg (might be a symlink) |
|
823 | 823 | continue; |
|
824 | 824 | } else if metadata.is_dir() { |
|
825 | 825 | // A .hg sub-directory at another location means a subrepo, |
|
826 | 826 | // skip it entirely. |
|
827 | 827 | return Ok(Vec::new()); |
|
828 | 828 | } |
|
829 | 829 | } |
|
830 | 830 | let full_path = if at_cwd { |
|
831 | 831 | file_name.clone().into() |
|
832 | 832 | } else { |
|
833 | 833 | entry.path() |
|
834 | 834 | }; |
|
835 | 835 | let base_name = get_bytes_from_os_string(file_name).into(); |
|
836 | 836 | results.push(DirEntry { |
|
837 | 837 | base_name, |
|
838 | 838 | full_path, |
|
839 | 839 | metadata, |
|
840 | 840 | }) |
|
841 | 841 | } |
|
842 | 842 | Ok(results) |
|
843 | 843 | } |
|
844 | 844 | } |
|
845 | 845 | |
|
846 | 846 | /// Return the `mtime` of a temporary file newly-created in the `.hg` directory |
|
847 | 847 | /// of the give repository. |
|
848 | 848 | /// |
|
849 | 849 | /// This is similar to `SystemTime::now()`, with the result truncated to the |
|
850 | 850 | /// same time resolution as other files’ modification times. Using `.hg` |
|
851 | 851 | /// instead of the system’s default temporary directory (such as `/tmp`) makes |
|
852 | 852 | /// it more likely the temporary file is in the same disk partition as contents |
|
853 | 853 | /// of the working directory, which can matter since different filesystems may |
|
854 | 854 | /// store timestamps with different resolutions. |
|
855 | 855 | /// |
|
856 | 856 | /// This may fail, typically if we lack write permissions. In that case we |
|
857 | 857 | /// should continue the `status()` algoritm anyway and consider the current |
|
858 | 858 | /// date/time to be unknown. |
|
859 | 859 | fn filesystem_now(repo_root: &Path) -> Result<SystemTime, io::Error> { |
|
860 | 860 | tempfile::tempfile_in(repo_root.join(".hg"))? |
|
861 | 861 | .metadata()? |
|
862 | 862 | .modified() |
|
863 | 863 | } |
@@ -1,13 +1,12 b'' | |||
|
1 | 1 | skip ignored directories if -i or --all not specified |
|
2 | 2 | |
|
3 | 3 | $ hg init t |
|
4 | 4 | $ cd t |
|
5 | 5 | $ echo 'ignored' > .hgignore |
|
6 | 6 | $ hg ci -qA -m init -d'2 0' |
|
7 | 7 | $ mkdir ignored |
|
8 | ||
|
9 | The better behavior here is the non-rust behavior, which is to keep | |
|
10 | the directory and only delete it when -i or --all is given. | |
|
11 | ||
|
8 | $ ls | |
|
9 | ignored | |
|
12 | 10 | $ hg purge -v --no-confirm |
|
13 | removing directory ignored (known-bad-output rust !) | |
|
11 | $ ls | |
|
12 | ignored |
General Comments 0
You need to be logged in to leave comments.
Login now