Show More
@@ -1,149 +1,149 b'' | |||
|
1 | 1 | // status.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 | //! Rust implementation of dirstate.status (dirstate.py). |
|
9 | 9 | //! It is currently missing a lot of functionality compared to the Python one |
|
10 | 10 | //! and will only be triggered in narrow cases. |
|
11 | 11 | |
|
12 | 12 | use crate::dirstate::entry::TruncatedTimestamp; |
|
13 | 13 | use crate::dirstate_tree::on_disk::DirstateV2ParseError; |
|
14 | 14 | use crate::{ |
|
15 | 15 | utils::hg_path::{HgPath, HgPathError}, |
|
16 | 16 | PatternError, |
|
17 | 17 | }; |
|
18 | 18 | |
|
19 | 19 | use std::{borrow::Cow, fmt}; |
|
20 | 20 | |
|
21 | 21 | /// Wrong type of file from a `BadMatch` |
|
22 | 22 | /// Note: a lot of those don't exist on all platforms. |
|
23 | #[derive(Debug, Copy, Clone)] | |
|
23 | #[derive(Debug, Copy, Clone, PartialEq, Eq)] | |
|
24 | 24 | pub enum BadType { |
|
25 | 25 | CharacterDevice, |
|
26 | 26 | BlockDevice, |
|
27 | 27 | FIFO, |
|
28 | 28 | Socket, |
|
29 | 29 | Directory, |
|
30 | 30 | Unknown, |
|
31 | 31 | } |
|
32 | 32 | |
|
33 | 33 | impl fmt::Display for BadType { |
|
34 | 34 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
|
35 | 35 | f.write_str(match self { |
|
36 | 36 | BadType::CharacterDevice => "character device", |
|
37 | 37 | BadType::BlockDevice => "block device", |
|
38 | 38 | BadType::FIFO => "fifo", |
|
39 | 39 | BadType::Socket => "socket", |
|
40 | 40 | BadType::Directory => "directory", |
|
41 | 41 | BadType::Unknown => "unknown", |
|
42 | 42 | }) |
|
43 | 43 | } |
|
44 | 44 | } |
|
45 | 45 | |
|
46 | 46 | /// Was explicitly matched but cannot be found/accessed |
|
47 | 47 | #[derive(Debug, Copy, Clone)] |
|
48 | 48 | pub enum BadMatch { |
|
49 | 49 | OsError(i32), |
|
50 | 50 | BadType(BadType), |
|
51 | 51 | } |
|
52 | 52 | |
|
53 | 53 | /// `Box<dyn Trait>` is syntactic sugar for `Box<dyn Trait + 'static>`, so add |
|
54 | 54 | /// an explicit lifetime here to not fight `'static` bounds "out of nowhere". |
|
55 | 55 | pub type IgnoreFnType<'a> = |
|
56 | 56 | Box<dyn for<'r> Fn(&'r HgPath) -> bool + Sync + 'a>; |
|
57 | 57 | |
|
58 | 58 | /// We have a good mix of owned (from directory traversal) and borrowed (from |
|
59 | 59 | /// the dirstate/explicit) paths, this comes up a lot. |
|
60 | 60 | pub type HgPathCow<'a> = Cow<'a, HgPath>; |
|
61 | 61 | |
|
62 | 62 | #[derive(Debug, Copy, Clone)] |
|
63 | 63 | pub struct StatusOptions { |
|
64 | 64 | /// Whether we are on a filesystem with UNIX-like exec flags |
|
65 | 65 | pub check_exec: bool, |
|
66 | 66 | pub list_clean: bool, |
|
67 | 67 | pub list_unknown: bool, |
|
68 | 68 | pub list_ignored: bool, |
|
69 | 69 | /// Whether to populate `StatusPath::copy_source` |
|
70 | 70 | pub list_copies: bool, |
|
71 | 71 | /// Whether to collect traversed dirs for applying a callback later. |
|
72 | 72 | /// Used by `hg purge` for example. |
|
73 | 73 | pub collect_traversed_dirs: bool, |
|
74 | 74 | } |
|
75 | 75 | |
|
76 | 76 | #[derive(Default)] |
|
77 | 77 | pub struct DirstateStatus<'a> { |
|
78 | 78 | /// The current time at the start of the `status()` algorithm, as measured |
|
79 | 79 | /// and possibly truncated by the filesystem. |
|
80 | 80 | pub filesystem_time_at_status_start: Option<TruncatedTimestamp>, |
|
81 | 81 | |
|
82 | 82 | /// Tracked files whose contents have changed since the parent revision |
|
83 | 83 | pub modified: Vec<StatusPath<'a>>, |
|
84 | 84 | |
|
85 | 85 | /// Newly-tracked files that were not present in the parent |
|
86 | 86 | pub added: Vec<StatusPath<'a>>, |
|
87 | 87 | |
|
88 | 88 | /// Previously-tracked files that have been (re)moved with an hg command |
|
89 | 89 | pub removed: Vec<StatusPath<'a>>, |
|
90 | 90 | |
|
91 | 91 | /// (Still) tracked files that are missing, (re)moved with an non-hg |
|
92 | 92 | /// command |
|
93 | 93 | pub deleted: Vec<StatusPath<'a>>, |
|
94 | 94 | |
|
95 | 95 | /// Tracked files that are up to date with the parent. |
|
96 | 96 | /// Only pupulated if `StatusOptions::list_clean` is true. |
|
97 | 97 | pub clean: Vec<StatusPath<'a>>, |
|
98 | 98 | |
|
99 | 99 | /// Files in the working directory that are ignored with `.hgignore`. |
|
100 | 100 | /// Only pupulated if `StatusOptions::list_ignored` is true. |
|
101 | 101 | pub ignored: Vec<StatusPath<'a>>, |
|
102 | 102 | |
|
103 | 103 | /// Files in the working directory that are neither tracked nor ignored. |
|
104 | 104 | /// Only pupulated if `StatusOptions::list_unknown` is true. |
|
105 | 105 | pub unknown: Vec<StatusPath<'a>>, |
|
106 | 106 | |
|
107 | 107 | /// Was explicitly matched but cannot be found/accessed |
|
108 | 108 | pub bad: Vec<(HgPathCow<'a>, BadMatch)>, |
|
109 | 109 | |
|
110 | 110 | /// Either clean or modified, but we can’t tell from filesystem metadata |
|
111 | 111 | /// alone. The file contents need to be read and compared with that in |
|
112 | 112 | /// the parent. |
|
113 | 113 | pub unsure: Vec<StatusPath<'a>>, |
|
114 | 114 | |
|
115 | 115 | /// Only filled if `collect_traversed_dirs` is `true` |
|
116 | 116 | pub traversed: Vec<HgPathCow<'a>>, |
|
117 | 117 | |
|
118 | 118 | /// Whether `status()` made changed to the `DirstateMap` that should be |
|
119 | 119 | /// written back to disk |
|
120 | 120 | pub dirty: bool, |
|
121 | 121 | } |
|
122 | 122 | |
|
123 | 123 | #[derive(Clone, PartialEq, Eq, PartialOrd, Ord)] |
|
124 | 124 | pub struct StatusPath<'a> { |
|
125 | 125 | pub path: HgPathCow<'a>, |
|
126 | 126 | pub copy_source: Option<HgPathCow<'a>>, |
|
127 | 127 | } |
|
128 | 128 | |
|
129 | 129 | #[derive(Debug, derive_more::From)] |
|
130 | 130 | pub enum StatusError { |
|
131 | 131 | /// An invalid path that cannot be represented in Mercurial was found |
|
132 | 132 | Path(HgPathError), |
|
133 | 133 | /// An invalid "ignore" pattern was found |
|
134 | 134 | Pattern(PatternError), |
|
135 | 135 | /// Corrupted dirstate |
|
136 | 136 | DirstateV2ParseError(DirstateV2ParseError), |
|
137 | 137 | } |
|
138 | 138 | |
|
139 | 139 | impl fmt::Display for StatusError { |
|
140 | 140 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
|
141 | 141 | match self { |
|
142 | 142 | StatusError::Path(error) => error.fmt(f), |
|
143 | 143 | StatusError::Pattern(error) => error.fmt(f), |
|
144 | 144 | StatusError::DirstateV2ParseError(_) => { |
|
145 | 145 | f.write_str("dirstate-v2 parse error") |
|
146 | 146 | } |
|
147 | 147 | } |
|
148 | 148 | } |
|
149 | 149 | } |
@@ -1,996 +1,1013 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_bytes_from_path; |
|
14 | 14 | use crate::utils::files::get_path_from_bytes; |
|
15 | 15 | use crate::utils::hg_path::HgPath; |
|
16 | 16 | use crate::BadMatch; |
|
17 | use crate::BadType; | |
|
17 | 18 | use crate::DirstateStatus; |
|
18 | 19 | use crate::HgPathCow; |
|
19 | 20 | use crate::PatternFileWarning; |
|
20 | 21 | use crate::StatusError; |
|
21 | 22 | use crate::StatusOptions; |
|
22 | 23 | use once_cell::sync::OnceCell; |
|
23 | 24 | use rayon::prelude::*; |
|
24 | 25 | use sha1::{Digest, Sha1}; |
|
25 | 26 | use std::borrow::Cow; |
|
26 | 27 | use std::io; |
|
28 | use std::os::unix::prelude::FileTypeExt; | |
|
27 | 29 | use std::path::Path; |
|
28 | 30 | use std::path::PathBuf; |
|
29 | 31 | use std::sync::Mutex; |
|
30 | 32 | use std::time::SystemTime; |
|
31 | 33 | |
|
32 | 34 | /// Returns the status of the working directory compared to its parent |
|
33 | 35 | /// changeset. |
|
34 | 36 | /// |
|
35 | 37 | /// This algorithm is based on traversing the filesystem tree (`fs` in function |
|
36 | 38 | /// and variable names) and dirstate tree at the same time. The core of this |
|
37 | 39 | /// traversal is the recursive `traverse_fs_directory_and_dirstate` function |
|
38 | 40 | /// and its use of `itertools::merge_join_by`. When reaching a path that only |
|
39 | 41 | /// exists in one of the two trees, depending on information requested by |
|
40 | 42 | /// `options` we may need to traverse the remaining subtree. |
|
41 | 43 | #[logging_timer::time("trace")] |
|
42 | 44 | pub fn status<'dirstate>( |
|
43 | 45 | dmap: &'dirstate mut DirstateMap, |
|
44 | 46 | matcher: &(dyn Matcher + Sync), |
|
45 | 47 | root_dir: PathBuf, |
|
46 | 48 | ignore_files: Vec<PathBuf>, |
|
47 | 49 | options: StatusOptions, |
|
48 | 50 | ) -> Result<(DirstateStatus<'dirstate>, Vec<PatternFileWarning>), StatusError> |
|
49 | 51 | { |
|
50 | 52 | // Also cap for a Python caller of this function, but don't complain if |
|
51 | 53 | // the global threadpool has already been set since this code path is also |
|
52 | 54 | // being used by `rhg`, which calls this early. |
|
53 | 55 | let _ = crate::utils::cap_default_rayon_threads(); |
|
54 | 56 | |
|
55 | 57 | let (ignore_fn, warnings, patterns_changed): (IgnoreFnType, _, _) = |
|
56 | 58 | if options.list_ignored || options.list_unknown { |
|
57 | 59 | let (ignore_fn, warnings, changed) = match dmap.dirstate_version { |
|
58 | 60 | DirstateVersion::V1 => { |
|
59 | 61 | let (ignore_fn, warnings) = get_ignore_function( |
|
60 | 62 | ignore_files, |
|
61 | 63 | &root_dir, |
|
62 | 64 | &mut |_source, _pattern_bytes| {}, |
|
63 | 65 | )?; |
|
64 | 66 | (ignore_fn, warnings, None) |
|
65 | 67 | } |
|
66 | 68 | DirstateVersion::V2 => { |
|
67 | 69 | let mut hasher = Sha1::new(); |
|
68 | 70 | let (ignore_fn, warnings) = get_ignore_function( |
|
69 | 71 | ignore_files, |
|
70 | 72 | &root_dir, |
|
71 | 73 | &mut |source, pattern_bytes| { |
|
72 | 74 | // If inside the repo, use the relative version to |
|
73 | 75 | // make it deterministic inside tests. |
|
74 | 76 | // The performance hit should be negligible. |
|
75 | 77 | let source = source |
|
76 | 78 | .strip_prefix(&root_dir) |
|
77 | 79 | .unwrap_or(source); |
|
78 | 80 | let source = get_bytes_from_path(source); |
|
79 | 81 | |
|
80 | 82 | let mut subhasher = Sha1::new(); |
|
81 | 83 | subhasher.update(pattern_bytes); |
|
82 | 84 | let patterns_hash = subhasher.finalize(); |
|
83 | 85 | |
|
84 | 86 | hasher.update(source); |
|
85 | 87 | hasher.update(b" "); |
|
86 | 88 | hasher.update(patterns_hash); |
|
87 | 89 | hasher.update(b"\n"); |
|
88 | 90 | }, |
|
89 | 91 | )?; |
|
90 | 92 | let new_hash = *hasher.finalize().as_ref(); |
|
91 | 93 | let changed = new_hash != dmap.ignore_patterns_hash; |
|
92 | 94 | dmap.ignore_patterns_hash = new_hash; |
|
93 | 95 | (ignore_fn, warnings, Some(changed)) |
|
94 | 96 | } |
|
95 | 97 | }; |
|
96 | 98 | (ignore_fn, warnings, changed) |
|
97 | 99 | } else { |
|
98 | 100 | (Box::new(|&_| true), vec![], None) |
|
99 | 101 | }; |
|
100 | 102 | |
|
101 | 103 | let filesystem_time_at_status_start = |
|
102 | 104 | filesystem_now(&root_dir).ok().map(TruncatedTimestamp::from); |
|
103 | 105 | |
|
104 | 106 | // If the repository is under the current directory, prefer using a |
|
105 | 107 | // relative path, so the kernel needs to traverse fewer directory in every |
|
106 | 108 | // call to `read_dir` or `symlink_metadata`. |
|
107 | 109 | // This is effective in the common case where the current directory is the |
|
108 | 110 | // repository root. |
|
109 | 111 | |
|
110 | 112 | // TODO: Better yet would be to use libc functions like `openat` and |
|
111 | 113 | // `fstatat` to remove such repeated traversals entirely, but the standard |
|
112 | 114 | // library does not provide APIs based on those. |
|
113 | 115 | // Maybe with a crate like https://crates.io/crates/openat instead? |
|
114 | 116 | let root_dir = if let Some(relative) = std::env::current_dir() |
|
115 | 117 | .ok() |
|
116 | 118 | .and_then(|cwd| root_dir.strip_prefix(cwd).ok()) |
|
117 | 119 | { |
|
118 | 120 | relative |
|
119 | 121 | } else { |
|
120 | 122 | &root_dir |
|
121 | 123 | }; |
|
122 | 124 | |
|
123 | 125 | let outcome = DirstateStatus { |
|
124 | 126 | filesystem_time_at_status_start, |
|
125 | 127 | ..Default::default() |
|
126 | 128 | }; |
|
127 | 129 | let common = StatusCommon { |
|
128 | 130 | dmap, |
|
129 | 131 | options, |
|
130 | 132 | matcher, |
|
131 | 133 | ignore_fn, |
|
132 | 134 | outcome: Mutex::new(outcome), |
|
133 | 135 | ignore_patterns_have_changed: patterns_changed, |
|
134 | 136 | new_cacheable_directories: Default::default(), |
|
135 | 137 | outdated_cached_directories: Default::default(), |
|
136 | 138 | filesystem_time_at_status_start, |
|
137 | 139 | }; |
|
138 | 140 | let is_at_repo_root = true; |
|
139 | 141 | let hg_path = &BorrowedPath::OnDisk(HgPath::new("")); |
|
140 | 142 | let has_ignored_ancestor = HasIgnoredAncestor::create(None, hg_path); |
|
141 | 143 | let root_cached_mtime = None; |
|
142 | 144 | // If the path we have for the repository root is a symlink, do follow it. |
|
143 | 145 | // (As opposed to symlinks within the working directory which are not |
|
144 | 146 | // followed, using `std::fs::symlink_metadata`.) |
|
145 | 147 | common.traverse_fs_directory_and_dirstate( |
|
146 | 148 | &has_ignored_ancestor, |
|
147 | 149 | dmap.root.as_ref(), |
|
148 | 150 | hg_path, |
|
149 | 151 | &DirEntry { |
|
150 | 152 | hg_path: Cow::Borrowed(HgPath::new(b"")), |
|
151 | 153 | fs_path: Cow::Borrowed(root_dir), |
|
152 | 154 | symlink_metadata: None, |
|
153 | 155 | file_type: FakeFileType::Directory, |
|
154 | 156 | }, |
|
155 | 157 | root_cached_mtime, |
|
156 | 158 | is_at_repo_root, |
|
157 | 159 | )?; |
|
158 | 160 | let mut outcome = common.outcome.into_inner().unwrap(); |
|
159 | 161 | let new_cacheable = common.new_cacheable_directories.into_inner().unwrap(); |
|
160 | 162 | let outdated = common.outdated_cached_directories.into_inner().unwrap(); |
|
161 | 163 | |
|
162 | 164 | outcome.dirty = common.ignore_patterns_have_changed == Some(true) |
|
163 | 165 | || !outdated.is_empty() |
|
164 | 166 | || (!new_cacheable.is_empty() |
|
165 | 167 | && dmap.dirstate_version == DirstateVersion::V2); |
|
166 | 168 | |
|
167 | 169 | // Remove outdated mtimes before adding new mtimes, in case a given |
|
168 | 170 | // directory is both |
|
169 | 171 | for path in &outdated { |
|
170 | 172 | dmap.clear_cached_mtime(path)?; |
|
171 | 173 | } |
|
172 | 174 | for (path, mtime) in &new_cacheable { |
|
173 | 175 | dmap.set_cached_mtime(path, *mtime)?; |
|
174 | 176 | } |
|
175 | 177 | |
|
176 | 178 | Ok((outcome, warnings)) |
|
177 | 179 | } |
|
178 | 180 | |
|
179 | 181 | /// Bag of random things needed by various parts of the algorithm. Reduces the |
|
180 | 182 | /// number of parameters passed to functions. |
|
181 | 183 | struct StatusCommon<'a, 'tree, 'on_disk: 'tree> { |
|
182 | 184 | dmap: &'tree DirstateMap<'on_disk>, |
|
183 | 185 | options: StatusOptions, |
|
184 | 186 | matcher: &'a (dyn Matcher + Sync), |
|
185 | 187 | ignore_fn: IgnoreFnType<'a>, |
|
186 | 188 | outcome: Mutex<DirstateStatus<'on_disk>>, |
|
187 | 189 | /// New timestamps of directories to be used for caching their readdirs |
|
188 | 190 | new_cacheable_directories: |
|
189 | 191 | Mutex<Vec<(Cow<'on_disk, HgPath>, TruncatedTimestamp)>>, |
|
190 | 192 | /// Used to invalidate the readdir cache of directories |
|
191 | 193 | outdated_cached_directories: Mutex<Vec<Cow<'on_disk, HgPath>>>, |
|
192 | 194 | |
|
193 | 195 | /// Whether ignore files like `.hgignore` have changed since the previous |
|
194 | 196 | /// time a `status()` call wrote their hash to the dirstate. `None` means |
|
195 | 197 | /// we don’t know as this run doesn’t list either ignored or uknown files |
|
196 | 198 | /// and therefore isn’t reading `.hgignore`. |
|
197 | 199 | ignore_patterns_have_changed: Option<bool>, |
|
198 | 200 | |
|
199 | 201 | /// The current time at the start of the `status()` algorithm, as measured |
|
200 | 202 | /// and possibly truncated by the filesystem. |
|
201 | 203 | filesystem_time_at_status_start: Option<TruncatedTimestamp>, |
|
202 | 204 | } |
|
203 | 205 | |
|
204 | 206 | enum Outcome { |
|
205 | 207 | Modified, |
|
206 | 208 | Added, |
|
207 | 209 | Removed, |
|
208 | 210 | Deleted, |
|
209 | 211 | Clean, |
|
210 | 212 | Ignored, |
|
211 | 213 | Unknown, |
|
212 | 214 | Unsure, |
|
213 | 215 | } |
|
214 | 216 | |
|
215 | 217 | /// Lazy computation of whether a given path has a hgignored |
|
216 | 218 | /// ancestor. |
|
217 | 219 | struct HasIgnoredAncestor<'a> { |
|
218 | 220 | /// `path` and `parent` constitute the inputs to the computation, |
|
219 | 221 | /// `cache` stores the outcome. |
|
220 | 222 | path: &'a HgPath, |
|
221 | 223 | parent: Option<&'a HasIgnoredAncestor<'a>>, |
|
222 | 224 | cache: OnceCell<bool>, |
|
223 | 225 | } |
|
224 | 226 | |
|
225 | 227 | impl<'a> HasIgnoredAncestor<'a> { |
|
226 | 228 | fn create( |
|
227 | 229 | parent: Option<&'a HasIgnoredAncestor<'a>>, |
|
228 | 230 | path: &'a HgPath, |
|
229 | 231 | ) -> HasIgnoredAncestor<'a> { |
|
230 | 232 | Self { |
|
231 | 233 | path, |
|
232 | 234 | parent, |
|
233 | 235 | cache: OnceCell::new(), |
|
234 | 236 | } |
|
235 | 237 | } |
|
236 | 238 | |
|
237 | 239 | fn force<'b>(&self, ignore_fn: &IgnoreFnType<'b>) -> bool { |
|
238 | 240 | match self.parent { |
|
239 | 241 | None => false, |
|
240 | 242 | Some(parent) => { |
|
241 | 243 | *(self.cache.get_or_init(|| { |
|
242 | 244 | parent.force(ignore_fn) || ignore_fn(self.path) |
|
243 | 245 | })) |
|
244 | 246 | } |
|
245 | 247 | } |
|
246 | 248 | } |
|
247 | 249 | } |
|
248 | 250 | |
|
249 | 251 | impl<'a, 'tree, 'on_disk> StatusCommon<'a, 'tree, 'on_disk> { |
|
250 | 252 | fn push_outcome( |
|
251 | 253 | &self, |
|
252 | 254 | which: Outcome, |
|
253 | 255 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
254 | 256 | ) -> Result<(), DirstateV2ParseError> { |
|
255 | 257 | let path = dirstate_node |
|
256 | 258 | .full_path_borrowed(self.dmap.on_disk)? |
|
257 | 259 | .detach_from_tree(); |
|
258 | 260 | let copy_source = if self.options.list_copies { |
|
259 | 261 | dirstate_node |
|
260 | 262 | .copy_source_borrowed(self.dmap.on_disk)? |
|
261 | 263 | .map(|source| source.detach_from_tree()) |
|
262 | 264 | } else { |
|
263 | 265 | None |
|
264 | 266 | }; |
|
265 | 267 | self.push_outcome_common(which, path, copy_source); |
|
266 | 268 | Ok(()) |
|
267 | 269 | } |
|
268 | 270 | |
|
269 | 271 | fn push_outcome_without_copy_source( |
|
270 | 272 | &self, |
|
271 | 273 | which: Outcome, |
|
272 | 274 | path: &BorrowedPath<'_, 'on_disk>, |
|
273 | 275 | ) { |
|
274 | 276 | self.push_outcome_common(which, path.detach_from_tree(), None) |
|
275 | 277 | } |
|
276 | 278 | |
|
277 | 279 | fn push_outcome_common( |
|
278 | 280 | &self, |
|
279 | 281 | which: Outcome, |
|
280 | 282 | path: HgPathCow<'on_disk>, |
|
281 | 283 | copy_source: Option<HgPathCow<'on_disk>>, |
|
282 | 284 | ) { |
|
283 | 285 | let mut outcome = self.outcome.lock().unwrap(); |
|
284 | 286 | let vec = match which { |
|
285 | 287 | Outcome::Modified => &mut outcome.modified, |
|
286 | 288 | Outcome::Added => &mut outcome.added, |
|
287 | 289 | Outcome::Removed => &mut outcome.removed, |
|
288 | 290 | Outcome::Deleted => &mut outcome.deleted, |
|
289 | 291 | Outcome::Clean => &mut outcome.clean, |
|
290 | 292 | Outcome::Ignored => &mut outcome.ignored, |
|
291 | 293 | Outcome::Unknown => &mut outcome.unknown, |
|
292 | 294 | Outcome::Unsure => &mut outcome.unsure, |
|
293 | 295 | }; |
|
294 | 296 | vec.push(StatusPath { path, copy_source }); |
|
295 | 297 | } |
|
296 | 298 | |
|
297 | 299 | fn read_dir( |
|
298 | 300 | &self, |
|
299 | 301 | hg_path: &HgPath, |
|
300 | 302 | fs_path: &Path, |
|
301 | 303 | is_at_repo_root: bool, |
|
302 | 304 | ) -> Result<Vec<DirEntry>, ()> { |
|
303 | 305 | DirEntry::read_dir(fs_path, is_at_repo_root) |
|
304 | 306 | .map_err(|error| self.io_error(error, hg_path)) |
|
305 | 307 | } |
|
306 | 308 | |
|
307 | 309 | fn io_error(&self, error: std::io::Error, hg_path: &HgPath) { |
|
308 | 310 | let errno = error.raw_os_error().expect("expected real OS error"); |
|
309 | 311 | self.outcome |
|
310 | 312 | .lock() |
|
311 | 313 | .unwrap() |
|
312 | 314 | .bad |
|
313 | 315 | .push((hg_path.to_owned().into(), BadMatch::OsError(errno))) |
|
314 | 316 | } |
|
315 | 317 | |
|
316 | 318 | fn check_for_outdated_directory_cache( |
|
317 | 319 | &self, |
|
318 | 320 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
319 | 321 | ) -> Result<bool, DirstateV2ParseError> { |
|
320 | 322 | if self.ignore_patterns_have_changed == Some(true) |
|
321 | 323 | && dirstate_node.cached_directory_mtime()?.is_some() |
|
322 | 324 | { |
|
323 | 325 | self.outdated_cached_directories.lock().unwrap().push( |
|
324 | 326 | dirstate_node |
|
325 | 327 | .full_path_borrowed(self.dmap.on_disk)? |
|
326 | 328 | .detach_from_tree(), |
|
327 | 329 | ); |
|
328 | 330 | return Ok(true); |
|
329 | 331 | } |
|
330 | 332 | Ok(false) |
|
331 | 333 | } |
|
332 | 334 | |
|
333 | 335 | /// If this returns true, we can get accurate results by only using |
|
334 | 336 | /// `symlink_metadata` for child nodes that exist in the dirstate and don’t |
|
335 | 337 | /// need to call `read_dir`. |
|
336 | 338 | fn can_skip_fs_readdir( |
|
337 | 339 | &self, |
|
338 | 340 | directory_entry: &DirEntry, |
|
339 | 341 | cached_directory_mtime: Option<TruncatedTimestamp>, |
|
340 | 342 | ) -> bool { |
|
341 | 343 | if !self.options.list_unknown && !self.options.list_ignored { |
|
342 | 344 | // All states that we care about listing have corresponding |
|
343 | 345 | // dirstate entries. |
|
344 | 346 | // This happens for example with `hg status -mard`. |
|
345 | 347 | return true; |
|
346 | 348 | } |
|
347 | 349 | if !self.options.list_ignored |
|
348 | 350 | && self.ignore_patterns_have_changed == Some(false) |
|
349 | 351 | { |
|
350 | 352 | if let Some(cached_mtime) = cached_directory_mtime { |
|
351 | 353 | // The dirstate contains a cached mtime for this directory, set |
|
352 | 354 | // by a previous run of the `status` algorithm which found this |
|
353 | 355 | // directory eligible for `read_dir` caching. |
|
354 | 356 | if let Ok(meta) = directory_entry.symlink_metadata() { |
|
355 | 357 | if cached_mtime |
|
356 | 358 | .likely_equal_to_mtime_of(&meta) |
|
357 | 359 | .unwrap_or(false) |
|
358 | 360 | { |
|
359 | 361 | // The mtime of that directory has not changed |
|
360 | 362 | // since then, which means that the results of |
|
361 | 363 | // `read_dir` should also be unchanged. |
|
362 | 364 | return true; |
|
363 | 365 | } |
|
364 | 366 | } |
|
365 | 367 | } |
|
366 | 368 | } |
|
367 | 369 | false |
|
368 | 370 | } |
|
369 | 371 | |
|
370 | 372 | /// Returns whether all child entries of the filesystem directory have a |
|
371 | 373 | /// corresponding dirstate node or are ignored. |
|
372 | 374 | fn traverse_fs_directory_and_dirstate<'ancestor>( |
|
373 | 375 | &self, |
|
374 | 376 | has_ignored_ancestor: &'ancestor HasIgnoredAncestor<'ancestor>, |
|
375 | 377 | dirstate_nodes: ChildNodesRef<'tree, 'on_disk>, |
|
376 | 378 | directory_hg_path: &BorrowedPath<'tree, 'on_disk>, |
|
377 | 379 | directory_entry: &DirEntry, |
|
378 | 380 | cached_directory_mtime: Option<TruncatedTimestamp>, |
|
379 | 381 | is_at_repo_root: bool, |
|
380 | 382 | ) -> Result<bool, DirstateV2ParseError> { |
|
381 | 383 | if self.can_skip_fs_readdir(directory_entry, cached_directory_mtime) { |
|
382 | 384 | dirstate_nodes |
|
383 | 385 | .par_iter() |
|
384 | 386 | .map(|dirstate_node| { |
|
385 | 387 | let fs_path = &directory_entry.fs_path; |
|
386 | 388 | let fs_path = fs_path.join(get_path_from_bytes( |
|
387 | 389 | dirstate_node.base_name(self.dmap.on_disk)?.as_bytes(), |
|
388 | 390 | )); |
|
389 | 391 | match std::fs::symlink_metadata(&fs_path) { |
|
390 | 392 | Ok(fs_metadata) => { |
|
391 | let file_type = | |
|
392 | match fs_metadata.file_type().try_into() { | |
|
393 | Ok(file_type) => file_type, | |
|
394 | Err(_) => return Ok(()), | |
|
395 | }; | |
|
393 | let file_type = fs_metadata.file_type().into(); | |
|
396 | 394 | let entry = DirEntry { |
|
397 | 395 | hg_path: Cow::Borrowed( |
|
398 | 396 | dirstate_node |
|
399 | 397 | .full_path(self.dmap.on_disk)?, |
|
400 | 398 | ), |
|
401 | 399 | fs_path: Cow::Borrowed(&fs_path), |
|
402 | 400 | symlink_metadata: Some(fs_metadata), |
|
403 | 401 | file_type, |
|
404 | 402 | }; |
|
405 | 403 | self.traverse_fs_and_dirstate( |
|
406 | 404 | &entry, |
|
407 | 405 | dirstate_node, |
|
408 | 406 | has_ignored_ancestor, |
|
409 | 407 | ) |
|
410 | 408 | } |
|
411 | 409 | Err(e) if e.kind() == std::io::ErrorKind::NotFound => { |
|
412 | 410 | self.traverse_dirstate_only(dirstate_node) |
|
413 | 411 | } |
|
414 | 412 | Err(error) => { |
|
415 | 413 | let hg_path = |
|
416 | 414 | dirstate_node.full_path(self.dmap.on_disk)?; |
|
417 | 415 | self.io_error(error, hg_path); |
|
418 | 416 | Ok(()) |
|
419 | 417 | } |
|
420 | 418 | } |
|
421 | 419 | }) |
|
422 | 420 | .collect::<Result<_, _>>()?; |
|
423 | 421 | |
|
424 | 422 | // We don’t know, so conservatively say this isn’t the case |
|
425 | 423 | let children_all_have_dirstate_node_or_are_ignored = false; |
|
426 | 424 | |
|
427 | 425 | return Ok(children_all_have_dirstate_node_or_are_ignored); |
|
428 | 426 | } |
|
429 | 427 | |
|
430 | 428 | let readdir_succeeded; |
|
431 | 429 | let mut fs_entries = if let Ok(entries) = self.read_dir( |
|
432 | 430 | directory_hg_path, |
|
433 | 431 | &directory_entry.fs_path, |
|
434 | 432 | is_at_repo_root, |
|
435 | 433 | ) { |
|
436 | 434 | readdir_succeeded = true; |
|
437 | 435 | entries |
|
438 | 436 | } else { |
|
439 | 437 | // Treat an unreadable directory (typically because of insufficient |
|
440 | 438 | // permissions) like an empty directory. `self.read_dir` has |
|
441 | 439 | // already called `self.io_error` so a warning will be emitted. |
|
442 | 440 | // We still need to remember that there was an error so that we |
|
443 | 441 | // know not to cache this result. |
|
444 | 442 | readdir_succeeded = false; |
|
445 | 443 | Vec::new() |
|
446 | 444 | }; |
|
447 | 445 | |
|
448 | 446 | // `merge_join_by` requires both its input iterators to be sorted: |
|
449 | 447 | |
|
450 | 448 | let dirstate_nodes = dirstate_nodes.sorted(); |
|
451 | 449 | // `sort_unstable_by_key` doesn’t allow keys borrowing from the value: |
|
452 | 450 | // https://github.com/rust-lang/rust/issues/34162 |
|
453 | 451 | fs_entries.sort_unstable_by(|e1, e2| e1.hg_path.cmp(&e2.hg_path)); |
|
454 | 452 | |
|
455 | 453 | // Propagate here any error that would happen inside the comparison |
|
456 | 454 | // callback below |
|
457 | 455 | for dirstate_node in &dirstate_nodes { |
|
458 | 456 | dirstate_node.base_name(self.dmap.on_disk)?; |
|
459 | 457 | } |
|
460 | 458 | itertools::merge_join_by( |
|
461 | 459 | dirstate_nodes, |
|
462 | 460 | &fs_entries, |
|
463 | 461 | |dirstate_node, fs_entry| { |
|
464 | 462 | // This `unwrap` never panics because we already propagated |
|
465 | 463 | // those errors above |
|
466 | 464 | dirstate_node |
|
467 | 465 | .base_name(self.dmap.on_disk) |
|
468 | 466 | .unwrap() |
|
469 | 467 | .cmp(&fs_entry.hg_path) |
|
470 | 468 | }, |
|
471 | 469 | ) |
|
472 | 470 | .par_bridge() |
|
473 | 471 | .map(|pair| { |
|
474 | 472 | use itertools::EitherOrBoth::*; |
|
475 | 473 | let has_dirstate_node_or_is_ignored = match pair { |
|
476 | 474 | Both(dirstate_node, fs_entry) => { |
|
477 | 475 | self.traverse_fs_and_dirstate( |
|
478 | 476 | fs_entry, |
|
479 | 477 | dirstate_node, |
|
480 | 478 | has_ignored_ancestor, |
|
481 | 479 | )?; |
|
482 | 480 | true |
|
483 | 481 | } |
|
484 | 482 | Left(dirstate_node) => { |
|
485 | 483 | self.traverse_dirstate_only(dirstate_node)?; |
|
486 | 484 | true |
|
487 | 485 | } |
|
488 | 486 | Right(fs_entry) => self.traverse_fs_only( |
|
489 | 487 | has_ignored_ancestor.force(&self.ignore_fn), |
|
490 | 488 | directory_hg_path, |
|
491 | 489 | fs_entry, |
|
492 | 490 | ), |
|
493 | 491 | }; |
|
494 | 492 | Ok(has_dirstate_node_or_is_ignored) |
|
495 | 493 | }) |
|
496 | 494 | .try_reduce(|| true, |a, b| Ok(a && b)) |
|
497 | 495 | .map(|res| res && readdir_succeeded) |
|
498 | 496 | } |
|
499 | 497 | |
|
500 | 498 | fn traverse_fs_and_dirstate<'ancestor>( |
|
501 | 499 | &self, |
|
502 | 500 | fs_entry: &DirEntry, |
|
503 | 501 | dirstate_node: NodeRef<'tree, 'on_disk>, |
|
504 | 502 | has_ignored_ancestor: &'ancestor HasIgnoredAncestor<'ancestor>, |
|
505 | 503 | ) -> Result<(), DirstateV2ParseError> { |
|
506 | 504 | let outdated_dircache = |
|
507 | 505 | self.check_for_outdated_directory_cache(&dirstate_node)?; |
|
508 | 506 | let hg_path = &dirstate_node.full_path_borrowed(self.dmap.on_disk)?; |
|
509 | 507 | let file_or_symlink = fs_entry.is_file() || fs_entry.is_symlink(); |
|
510 | 508 | if !file_or_symlink { |
|
511 | 509 | // If we previously had a file here, it was removed (with |
|
512 | 510 | // `hg rm` or similar) or deleted before it could be |
|
513 | 511 | // replaced by a directory or something else. |
|
514 | 512 | self.mark_removed_or_deleted_if_file(&dirstate_node)?; |
|
515 | 513 | } |
|
514 | if let Some(bad_type) = fs_entry.is_bad() { | |
|
515 | if self.matcher.exact_match(hg_path) { | |
|
516 | let path = dirstate_node.full_path(self.dmap.on_disk)?; | |
|
517 | self.outcome.lock().unwrap().bad.push(( | |
|
518 | path.to_owned().into(), | |
|
519 | BadMatch::BadType(bad_type), | |
|
520 | )) | |
|
521 | } | |
|
522 | } | |
|
516 | 523 | if fs_entry.is_dir() { |
|
517 | 524 | if self.options.collect_traversed_dirs { |
|
518 | 525 | self.outcome |
|
519 | 526 | .lock() |
|
520 | 527 | .unwrap() |
|
521 | 528 | .traversed |
|
522 | 529 | .push(hg_path.detach_from_tree()) |
|
523 | 530 | } |
|
524 | 531 | let is_ignored = HasIgnoredAncestor::create( |
|
525 | 532 | Some(has_ignored_ancestor), |
|
526 | 533 | hg_path, |
|
527 | 534 | ); |
|
528 | 535 | let is_at_repo_root = false; |
|
529 | 536 | let children_all_have_dirstate_node_or_are_ignored = self |
|
530 | 537 | .traverse_fs_directory_and_dirstate( |
|
531 | 538 | &is_ignored, |
|
532 | 539 | dirstate_node.children(self.dmap.on_disk)?, |
|
533 | 540 | hg_path, |
|
534 | 541 | fs_entry, |
|
535 | 542 | dirstate_node.cached_directory_mtime()?, |
|
536 | 543 | is_at_repo_root, |
|
537 | 544 | )?; |
|
538 | 545 | self.maybe_save_directory_mtime( |
|
539 | 546 | children_all_have_dirstate_node_or_are_ignored, |
|
540 | 547 | fs_entry, |
|
541 | 548 | dirstate_node, |
|
542 | 549 | outdated_dircache, |
|
543 | 550 | )? |
|
544 | 551 | } else { |
|
545 | 552 | if file_or_symlink && self.matcher.matches(hg_path) { |
|
546 | 553 | if let Some(entry) = dirstate_node.entry()? { |
|
547 | 554 | if !entry.any_tracked() { |
|
548 | 555 | // Forward-compat if we start tracking unknown/ignored |
|
549 | 556 | // files for caching reasons |
|
550 | 557 | self.mark_unknown_or_ignored( |
|
551 | 558 | has_ignored_ancestor.force(&self.ignore_fn), |
|
552 | 559 | hg_path, |
|
553 | 560 | ); |
|
554 | 561 | } |
|
555 | 562 | if entry.added() { |
|
556 | 563 | self.push_outcome(Outcome::Added, &dirstate_node)?; |
|
557 | 564 | } else if entry.removed() { |
|
558 | 565 | self.push_outcome(Outcome::Removed, &dirstate_node)?; |
|
559 | 566 | } else if entry.modified() { |
|
560 | 567 | self.push_outcome(Outcome::Modified, &dirstate_node)?; |
|
561 | 568 | } else { |
|
562 | 569 | self.handle_normal_file(&dirstate_node, fs_entry)?; |
|
563 | 570 | } |
|
564 | 571 | } else { |
|
565 | 572 | // `node.entry.is_none()` indicates a "directory" |
|
566 | 573 | // node, but the filesystem has a file |
|
567 | 574 | self.mark_unknown_or_ignored( |
|
568 | 575 | has_ignored_ancestor.force(&self.ignore_fn), |
|
569 | 576 | hg_path, |
|
570 | 577 | ); |
|
571 | 578 | } |
|
572 | 579 | } |
|
573 | 580 | |
|
574 | 581 | for child_node in dirstate_node.children(self.dmap.on_disk)?.iter() |
|
575 | 582 | { |
|
576 | 583 | self.traverse_dirstate_only(child_node)? |
|
577 | 584 | } |
|
578 | 585 | } |
|
579 | 586 | Ok(()) |
|
580 | 587 | } |
|
581 | 588 | |
|
582 | 589 | /// Save directory mtime if applicable. |
|
583 | 590 | /// |
|
584 | 591 | /// `outdated_directory_cache` is `true` if we've just invalidated the |
|
585 | 592 | /// cache for this directory in `check_for_outdated_directory_cache`, |
|
586 | 593 | /// which forces the update. |
|
587 | 594 | fn maybe_save_directory_mtime( |
|
588 | 595 | &self, |
|
589 | 596 | children_all_have_dirstate_node_or_are_ignored: bool, |
|
590 | 597 | directory_entry: &DirEntry, |
|
591 | 598 | dirstate_node: NodeRef<'tree, 'on_disk>, |
|
592 | 599 | outdated_directory_cache: bool, |
|
593 | 600 | ) -> Result<(), DirstateV2ParseError> { |
|
594 | 601 | if !children_all_have_dirstate_node_or_are_ignored { |
|
595 | 602 | return Ok(()); |
|
596 | 603 | } |
|
597 | 604 | // All filesystem directory entries from `read_dir` have a |
|
598 | 605 | // corresponding node in the dirstate, so we can reconstitute the |
|
599 | 606 | // names of those entries without calling `read_dir` again. |
|
600 | 607 | |
|
601 | 608 | // TODO: use let-else here and below when available: |
|
602 | 609 | // https://github.com/rust-lang/rust/issues/87335 |
|
603 | 610 | let status_start = if let Some(status_start) = |
|
604 | 611 | &self.filesystem_time_at_status_start |
|
605 | 612 | { |
|
606 | 613 | status_start |
|
607 | 614 | } else { |
|
608 | 615 | return Ok(()); |
|
609 | 616 | }; |
|
610 | 617 | |
|
611 | 618 | // Although the Rust standard library’s `SystemTime` type |
|
612 | 619 | // has nanosecond precision, the times reported for a |
|
613 | 620 | // directory’s (or file’s) modified time may have lower |
|
614 | 621 | // resolution based on the filesystem (for example ext3 |
|
615 | 622 | // only stores integer seconds), kernel (see |
|
616 | 623 | // https://stackoverflow.com/a/14393315/1162888), etc. |
|
617 | 624 | let metadata = match directory_entry.symlink_metadata() { |
|
618 | 625 | Ok(meta) => meta, |
|
619 | 626 | Err(_) => return Ok(()), |
|
620 | 627 | }; |
|
621 | 628 | |
|
622 | 629 | let directory_mtime = match TruncatedTimestamp::for_reliable_mtime_of( |
|
623 | 630 | &metadata, |
|
624 | 631 | status_start, |
|
625 | 632 | ) { |
|
626 | 633 | Ok(Some(directory_mtime)) => directory_mtime, |
|
627 | 634 | Ok(None) => { |
|
628 | 635 | // The directory was modified too recently, |
|
629 | 636 | // don’t cache its `read_dir` results. |
|
630 | 637 | // |
|
631 | 638 | // 1. A change to this directory (direct child was |
|
632 | 639 | // added or removed) cause its mtime to be set |
|
633 | 640 | // (possibly truncated) to `directory_mtime` |
|
634 | 641 | // 2. This `status` algorithm calls `read_dir` |
|
635 | 642 | // 3. An other change is made to the same directory is |
|
636 | 643 | // made so that calling `read_dir` agin would give |
|
637 | 644 | // different results, but soon enough after 1. that |
|
638 | 645 | // the mtime stays the same |
|
639 | 646 | // |
|
640 | 647 | // On a system where the time resolution poor, this |
|
641 | 648 | // scenario is not unlikely if all three steps are caused |
|
642 | 649 | // by the same script. |
|
643 | 650 | return Ok(()); |
|
644 | 651 | } |
|
645 | 652 | Err(_) => { |
|
646 | 653 | // OS/libc does not support mtime? |
|
647 | 654 | return Ok(()); |
|
648 | 655 | } |
|
649 | 656 | }; |
|
650 | 657 | // We’ve observed (through `status_start`) that time has |
|
651 | 658 | // “progressed” since `directory_mtime`, so any further |
|
652 | 659 | // change to this directory is extremely likely to cause a |
|
653 | 660 | // different mtime. |
|
654 | 661 | // |
|
655 | 662 | // Having the same mtime again is not entirely impossible |
|
656 | 663 | // since the system clock is not monotonous. It could jump |
|
657 | 664 | // backward to some point before `directory_mtime`, then a |
|
658 | 665 | // directory change could potentially happen during exactly |
|
659 | 666 | // the wrong tick. |
|
660 | 667 | // |
|
661 | 668 | // We deem this scenario (unlike the previous one) to be |
|
662 | 669 | // unlikely enough in practice. |
|
663 | 670 | |
|
664 | 671 | let is_up_to_date = if let Some(cached) = |
|
665 | 672 | dirstate_node.cached_directory_mtime()? |
|
666 | 673 | { |
|
667 | 674 | !outdated_directory_cache && cached.likely_equal(directory_mtime) |
|
668 | 675 | } else { |
|
669 | 676 | false |
|
670 | 677 | }; |
|
671 | 678 | if !is_up_to_date { |
|
672 | 679 | let hg_path = dirstate_node |
|
673 | 680 | .full_path_borrowed(self.dmap.on_disk)? |
|
674 | 681 | .detach_from_tree(); |
|
675 | 682 | self.new_cacheable_directories |
|
676 | 683 | .lock() |
|
677 | 684 | .unwrap() |
|
678 | 685 | .push((hg_path, directory_mtime)) |
|
679 | 686 | } |
|
680 | 687 | Ok(()) |
|
681 | 688 | } |
|
682 | 689 | |
|
683 | 690 | /// A file that is clean in the dirstate was found in the filesystem |
|
684 | 691 | fn handle_normal_file( |
|
685 | 692 | &self, |
|
686 | 693 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
687 | 694 | fs_entry: &DirEntry, |
|
688 | 695 | ) -> Result<(), DirstateV2ParseError> { |
|
689 | 696 | // Keep the low 31 bits |
|
690 | 697 | fn truncate_u64(value: u64) -> i32 { |
|
691 | 698 | (value & 0x7FFF_FFFF) as i32 |
|
692 | 699 | } |
|
693 | 700 | |
|
694 | 701 | let fs_metadata = match fs_entry.symlink_metadata() { |
|
695 | 702 | Ok(meta) => meta, |
|
696 | 703 | Err(_) => return Ok(()), |
|
697 | 704 | }; |
|
698 | 705 | |
|
699 | 706 | let entry = dirstate_node |
|
700 | 707 | .entry()? |
|
701 | 708 | .expect("handle_normal_file called with entry-less node"); |
|
702 | 709 | let mode_changed = |
|
703 | 710 | || self.options.check_exec && entry.mode_changed(&fs_metadata); |
|
704 | 711 | let size = entry.size(); |
|
705 | 712 | let size_changed = size != truncate_u64(fs_metadata.len()); |
|
706 | 713 | if size >= 0 && size_changed && fs_metadata.file_type().is_symlink() { |
|
707 | 714 | // issue6456: Size returned may be longer due to encryption |
|
708 | 715 | // on EXT-4 fscrypt. TODO maybe only do it on EXT4? |
|
709 | 716 | self.push_outcome(Outcome::Unsure, dirstate_node)? |
|
710 | 717 | } else if dirstate_node.has_copy_source() |
|
711 | 718 | || entry.is_from_other_parent() |
|
712 | 719 | || (size >= 0 && (size_changed || mode_changed())) |
|
713 | 720 | { |
|
714 | 721 | self.push_outcome(Outcome::Modified, dirstate_node)? |
|
715 | 722 | } else { |
|
716 | 723 | let mtime_looks_clean = if let Some(dirstate_mtime) = |
|
717 | 724 | entry.truncated_mtime() |
|
718 | 725 | { |
|
719 | 726 | let fs_mtime = TruncatedTimestamp::for_mtime_of(&fs_metadata) |
|
720 | 727 | .expect("OS/libc does not support mtime?"); |
|
721 | 728 | // There might be a change in the future if for example the |
|
722 | 729 | // internal clock become off while process run, but this is a |
|
723 | 730 | // case where the issues the user would face |
|
724 | 731 | // would be a lot worse and there is nothing we |
|
725 | 732 | // can really do. |
|
726 | 733 | fs_mtime.likely_equal(dirstate_mtime) |
|
727 | 734 | } else { |
|
728 | 735 | // No mtime in the dirstate entry |
|
729 | 736 | false |
|
730 | 737 | }; |
|
731 | 738 | if !mtime_looks_clean { |
|
732 | 739 | self.push_outcome(Outcome::Unsure, dirstate_node)? |
|
733 | 740 | } else if self.options.list_clean { |
|
734 | 741 | self.push_outcome(Outcome::Clean, dirstate_node)? |
|
735 | 742 | } |
|
736 | 743 | } |
|
737 | 744 | Ok(()) |
|
738 | 745 | } |
|
739 | 746 | |
|
740 | 747 | /// A node in the dirstate tree has no corresponding filesystem entry |
|
741 | 748 | fn traverse_dirstate_only( |
|
742 | 749 | &self, |
|
743 | 750 | dirstate_node: NodeRef<'tree, 'on_disk>, |
|
744 | 751 | ) -> Result<(), DirstateV2ParseError> { |
|
745 | 752 | self.check_for_outdated_directory_cache(&dirstate_node)?; |
|
746 | 753 | self.mark_removed_or_deleted_if_file(&dirstate_node)?; |
|
747 | 754 | dirstate_node |
|
748 | 755 | .children(self.dmap.on_disk)? |
|
749 | 756 | .par_iter() |
|
750 | 757 | .map(|child_node| self.traverse_dirstate_only(child_node)) |
|
751 | 758 | .collect() |
|
752 | 759 | } |
|
753 | 760 | |
|
754 | 761 | /// A node in the dirstate tree has no corresponding *file* on the |
|
755 | 762 | /// filesystem |
|
756 | 763 | /// |
|
757 | 764 | /// Does nothing on a "directory" node |
|
758 | 765 | fn mark_removed_or_deleted_if_file( |
|
759 | 766 | &self, |
|
760 | 767 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
761 | 768 | ) -> Result<(), DirstateV2ParseError> { |
|
762 | 769 | if let Some(entry) = dirstate_node.entry()? { |
|
763 | 770 | if !entry.any_tracked() { |
|
764 | 771 | // Future-compat for when we start storing ignored and unknown |
|
765 | 772 | // files for caching reasons |
|
766 | 773 | return Ok(()); |
|
767 | 774 | } |
|
768 | 775 | let path = dirstate_node.full_path(self.dmap.on_disk)?; |
|
769 | 776 | if self.matcher.matches(path) { |
|
770 | 777 | if entry.removed() { |
|
771 | 778 | self.push_outcome(Outcome::Removed, dirstate_node)? |
|
772 | 779 | } else { |
|
773 | 780 | self.push_outcome(Outcome::Deleted, dirstate_node)? |
|
774 | 781 | } |
|
775 | 782 | } |
|
776 | 783 | } |
|
777 | 784 | Ok(()) |
|
778 | 785 | } |
|
779 | 786 | |
|
780 | 787 | /// Something in the filesystem has no corresponding dirstate node |
|
781 | 788 | /// |
|
782 | 789 | /// Returns whether that path is ignored |
|
783 | 790 | fn traverse_fs_only( |
|
784 | 791 | &self, |
|
785 | 792 | has_ignored_ancestor: bool, |
|
786 | 793 | directory_hg_path: &HgPath, |
|
787 | 794 | fs_entry: &DirEntry, |
|
788 | 795 | ) -> bool { |
|
789 | 796 | let hg_path = directory_hg_path.join(&fs_entry.hg_path); |
|
790 | 797 | let file_or_symlink = fs_entry.is_file() || fs_entry.is_symlink(); |
|
791 | 798 | if fs_entry.is_dir() { |
|
792 | 799 | let is_ignored = |
|
793 | 800 | has_ignored_ancestor || (self.ignore_fn)(&hg_path); |
|
794 | 801 | let traverse_children = if is_ignored { |
|
795 | 802 | // Descendants of an ignored directory are all ignored |
|
796 | 803 | self.options.list_ignored |
|
797 | 804 | } else { |
|
798 | 805 | // Descendants of an unknown directory may be either unknown or |
|
799 | 806 | // ignored |
|
800 | 807 | self.options.list_unknown || self.options.list_ignored |
|
801 | 808 | }; |
|
802 | 809 | if traverse_children { |
|
803 | 810 | let is_at_repo_root = false; |
|
804 | 811 | if let Ok(children_fs_entries) = |
|
805 | 812 | self.read_dir(&hg_path, &fs_entry.fs_path, is_at_repo_root) |
|
806 | 813 | { |
|
807 | 814 | children_fs_entries.par_iter().for_each(|child_fs_entry| { |
|
808 | 815 | self.traverse_fs_only( |
|
809 | 816 | is_ignored, |
|
810 | 817 | &hg_path, |
|
811 | 818 | child_fs_entry, |
|
812 | 819 | ); |
|
813 | 820 | }) |
|
814 | 821 | } |
|
815 | 822 | if self.options.collect_traversed_dirs { |
|
816 | 823 | self.outcome.lock().unwrap().traversed.push(hg_path.into()) |
|
817 | 824 | } |
|
818 | 825 | } |
|
819 | 826 | is_ignored |
|
820 | 827 | } else if file_or_symlink { |
|
821 | 828 | if self.matcher.matches(&hg_path) { |
|
822 | 829 | self.mark_unknown_or_ignored( |
|
823 | 830 | has_ignored_ancestor, |
|
824 | 831 | &BorrowedPath::InMemory(&hg_path), |
|
825 | 832 | ) |
|
826 | 833 | } else { |
|
827 | 834 | // We haven’t computed whether this path is ignored. It |
|
828 | 835 | // might not be, and a future run of status might have a |
|
829 | 836 | // different matcher that matches it. So treat it as not |
|
830 | 837 | // ignored. That is, inhibit readdir caching of the parent |
|
831 | 838 | // directory. |
|
832 | 839 | false |
|
833 | 840 | } |
|
834 | 841 | } else { |
|
835 | 842 | // This is neither a directory, a plain file, or a symlink. |
|
836 | 843 | // Treat it like an ignored file. |
|
837 | 844 | true |
|
838 | 845 | } |
|
839 | 846 | } |
|
840 | 847 | |
|
841 | 848 | /// Returns whether that path is ignored |
|
842 | 849 | fn mark_unknown_or_ignored( |
|
843 | 850 | &self, |
|
844 | 851 | has_ignored_ancestor: bool, |
|
845 | 852 | hg_path: &BorrowedPath<'_, 'on_disk>, |
|
846 | 853 | ) -> bool { |
|
847 | 854 | let is_ignored = has_ignored_ancestor || (self.ignore_fn)(hg_path); |
|
848 | 855 | if is_ignored { |
|
849 | 856 | if self.options.list_ignored { |
|
850 | 857 | self.push_outcome_without_copy_source( |
|
851 | 858 | Outcome::Ignored, |
|
852 | 859 | hg_path, |
|
853 | 860 | ) |
|
854 | 861 | } |
|
855 | 862 | } else if self.options.list_unknown { |
|
856 | 863 | self.push_outcome_without_copy_source(Outcome::Unknown, hg_path) |
|
857 | 864 | } |
|
858 | 865 | is_ignored |
|
859 | 866 | } |
|
860 | 867 | } |
|
861 | 868 | |
|
862 | 869 | /// Since [`std::fs::FileType`] cannot be built directly, we emulate what we |
|
863 | 870 | /// care about. |
|
864 | 871 | #[derive(Copy, Clone, Debug, PartialEq, Eq)] |
|
865 | 872 | enum FakeFileType { |
|
866 | 873 | File, |
|
867 | 874 | Directory, |
|
868 | 875 | Symlink, |
|
876 | BadType(BadType), | |
|
869 | 877 | } |
|
870 | 878 | |
|
871 |
impl |
|
|
872 | type Error = (); | |
|
873 | ||
|
874 | fn try_from(f: std::fs::FileType) -> Result<Self, Self::Error> { | |
|
879 | impl From<std::fs::FileType> for FakeFileType { | |
|
880 | fn from(f: std::fs::FileType) -> Self { | |
|
875 | 881 | if f.is_dir() { |
|
876 |
|
|
|
882 | Self::Directory | |
|
877 | 883 | } else if f.is_file() { |
|
878 |
|
|
|
884 | Self::File | |
|
879 | 885 | } else if f.is_symlink() { |
|
880 |
|
|
|
886 | Self::Symlink | |
|
887 | } else if f.is_fifo() { | |
|
888 | Self::BadType(BadType::FIFO) | |
|
889 | } else if f.is_block_device() { | |
|
890 | Self::BadType(BadType::BlockDevice) | |
|
891 | } else if f.is_char_device() { | |
|
892 | Self::BadType(BadType::CharacterDevice) | |
|
893 | } else if f.is_socket() { | |
|
894 | Self::BadType(BadType::Socket) | |
|
881 | 895 | } else { |
|
882 | // Things like FIFO etc. | |
|
883 | Err(()) | |
|
896 | Self::BadType(BadType::Unknown) | |
|
884 | 897 | } |
|
885 | 898 | } |
|
886 | 899 | } |
|
887 | 900 | |
|
888 | 901 | struct DirEntry<'a> { |
|
889 | 902 | /// Path as stored in the dirstate, or just the filename for optimization. |
|
890 | 903 | hg_path: HgPathCow<'a>, |
|
891 | 904 | /// Filesystem path |
|
892 | 905 | fs_path: Cow<'a, Path>, |
|
893 | 906 | /// Lazily computed |
|
894 | 907 | symlink_metadata: Option<std::fs::Metadata>, |
|
895 | 908 | /// Already computed for ergonomics. |
|
896 | 909 | file_type: FakeFileType, |
|
897 | 910 | } |
|
898 | 911 | |
|
899 | 912 | impl<'a> DirEntry<'a> { |
|
900 | 913 | /// Returns **unsorted** entries in the given directory, with name, |
|
901 | 914 | /// metadata and file type. |
|
902 | 915 | /// |
|
903 | 916 | /// If a `.hg` sub-directory is encountered: |
|
904 | 917 | /// |
|
905 | 918 | /// * At the repository root, ignore that sub-directory |
|
906 | 919 | /// * Elsewhere, we’re listing the content of a sub-repo. Return an empty |
|
907 | 920 | /// list instead. |
|
908 | 921 | fn read_dir(path: &Path, is_at_repo_root: bool) -> io::Result<Vec<Self>> { |
|
909 | 922 | // `read_dir` returns a "not found" error for the empty path |
|
910 | 923 | let at_cwd = path == Path::new(""); |
|
911 | 924 | let read_dir_path = if at_cwd { Path::new(".") } else { path }; |
|
912 | 925 | let mut results = Vec::new(); |
|
913 | 926 | for entry in read_dir_path.read_dir()? { |
|
914 | 927 | let entry = entry?; |
|
915 | 928 | let file_type = match entry.file_type() { |
|
916 | 929 | Ok(v) => v, |
|
917 | 930 | Err(e) => { |
|
918 | 931 | // race with file deletion? |
|
919 | 932 | if e.kind() == std::io::ErrorKind::NotFound { |
|
920 | 933 | continue; |
|
921 | 934 | } else { |
|
922 | 935 | return Err(e); |
|
923 | 936 | } |
|
924 | 937 | } |
|
925 | 938 | }; |
|
926 | 939 | let file_name = entry.file_name(); |
|
927 | 940 | // FIXME don't do this when cached |
|
928 | 941 | if file_name == ".hg" { |
|
929 | 942 | if is_at_repo_root { |
|
930 | 943 | // Skip the repo’s own .hg (might be a symlink) |
|
931 | 944 | continue; |
|
932 | 945 | } else if file_type.is_dir() { |
|
933 | 946 | // A .hg sub-directory at another location means a subrepo, |
|
934 | 947 | // skip it entirely. |
|
935 | 948 | return Ok(Vec::new()); |
|
936 | 949 | } |
|
937 | 950 | } |
|
938 | 951 | let full_path = if at_cwd { |
|
939 | 952 | file_name.clone().into() |
|
940 | 953 | } else { |
|
941 | 954 | entry.path() |
|
942 | 955 | }; |
|
943 | 956 | let filename = |
|
944 | 957 | Cow::Owned(get_bytes_from_os_string(file_name).into()); |
|
945 |
let file_type = |
|
|
946 | Ok(file_type) => file_type, | |
|
947 | Err(_) => continue, | |
|
948 | }; | |
|
958 | let file_type = FakeFileType::from(file_type); | |
|
949 | 959 | results.push(DirEntry { |
|
950 | 960 | hg_path: filename, |
|
951 | 961 | fs_path: Cow::Owned(full_path.to_path_buf()), |
|
952 | 962 | symlink_metadata: None, |
|
953 | 963 | file_type, |
|
954 | 964 | }) |
|
955 | 965 | } |
|
956 | 966 | Ok(results) |
|
957 | 967 | } |
|
958 | 968 | |
|
959 | 969 | fn symlink_metadata(&self) -> Result<std::fs::Metadata, std::io::Error> { |
|
960 | 970 | match &self.symlink_metadata { |
|
961 | 971 | Some(meta) => Ok(meta.clone()), |
|
962 | 972 | None => std::fs::symlink_metadata(&self.fs_path), |
|
963 | 973 | } |
|
964 | 974 | } |
|
965 | 975 | |
|
966 | 976 | fn is_dir(&self) -> bool { |
|
967 | 977 | self.file_type == FakeFileType::Directory |
|
968 | 978 | } |
|
969 | 979 | |
|
970 | 980 | fn is_file(&self) -> bool { |
|
971 | 981 | self.file_type == FakeFileType::File |
|
972 | 982 | } |
|
973 | 983 | |
|
974 | 984 | fn is_symlink(&self) -> bool { |
|
975 | 985 | self.file_type == FakeFileType::Symlink |
|
976 | 986 | } |
|
987 | ||
|
988 | fn is_bad(&self) -> Option<BadType> { | |
|
989 | match self.file_type { | |
|
990 | FakeFileType::BadType(ty) => Some(ty), | |
|
991 | _ => None, | |
|
992 | } | |
|
993 | } | |
|
977 | 994 | } |
|
978 | 995 | |
|
979 | 996 | /// Return the `mtime` of a temporary file newly-created in the `.hg` directory |
|
980 | 997 | /// of the give repository. |
|
981 | 998 | /// |
|
982 | 999 | /// This is similar to `SystemTime::now()`, with the result truncated to the |
|
983 | 1000 | /// same time resolution as other files’ modification times. Using `.hg` |
|
984 | 1001 | /// instead of the system’s default temporary directory (such as `/tmp`) makes |
|
985 | 1002 | /// it more likely the temporary file is in the same disk partition as contents |
|
986 | 1003 | /// of the working directory, which can matter since different filesystems may |
|
987 | 1004 | /// store timestamps with different resolutions. |
|
988 | 1005 | /// |
|
989 | 1006 | /// This may fail, typically if we lack write permissions. In that case we |
|
990 | 1007 | /// should continue the `status()` algoritm anyway and consider the current |
|
991 | 1008 | /// date/time to be unknown. |
|
992 | 1009 | fn filesystem_now(repo_root: &Path) -> Result<SystemTime, io::Error> { |
|
993 | 1010 | tempfile::tempfile_in(repo_root.join(".hg"))? |
|
994 | 1011 | .metadata()? |
|
995 | 1012 | .modified() |
|
996 | 1013 | } |
General Comments 0
You need to be logged in to leave comments.
Login now