Show More
@@ -1,353 +1,359 b'' | |||
|
1 | 1 | use crate::errors::HgError; |
|
2 | 2 | use crate::revlog::Revision; |
|
3 | 3 | use crate::revlog::{Node, NodePrefix}; |
|
4 | 4 | use crate::revlog::{Revlog, RevlogEntry, RevlogError}; |
|
5 | 5 | use crate::utils::hg_path::HgPath; |
|
6 | 6 | use crate::vfs::Vfs; |
|
7 | use crate::UncheckedRevision; | |
|
7 | use crate::{Graph, GraphError, UncheckedRevision}; | |
|
8 | 8 | use itertools::Itertools; |
|
9 | 9 | use std::ascii::escape_default; |
|
10 | 10 | use std::borrow::Cow; |
|
11 | 11 | use std::fmt::{Debug, Formatter}; |
|
12 | 12 | |
|
13 | 13 | /// A specialized `Revlog` to work with changelog data format. |
|
14 | 14 | pub struct Changelog { |
|
15 | 15 | /// The generic `revlog` format. |
|
16 | 16 | pub(crate) revlog: Revlog, |
|
17 | 17 | } |
|
18 | 18 | |
|
19 | 19 | impl Changelog { |
|
20 | 20 | /// Open the `changelog` of a repository given by its root. |
|
21 | 21 | pub fn open(store_vfs: &Vfs, use_nodemap: bool) -> Result<Self, HgError> { |
|
22 | 22 | let revlog = |
|
23 | 23 | Revlog::open(store_vfs, "00changelog.i", None, use_nodemap)?; |
|
24 | 24 | Ok(Self { revlog }) |
|
25 | 25 | } |
|
26 | 26 | |
|
27 | 27 | /// Return the `ChangelogRevisionData` for the given node ID. |
|
28 | 28 | pub fn data_for_node( |
|
29 | 29 | &self, |
|
30 | 30 | node: NodePrefix, |
|
31 | 31 | ) -> Result<ChangelogRevisionData, RevlogError> { |
|
32 | 32 | let rev = self.revlog.rev_from_node(node)?; |
|
33 | 33 | self.entry_for_checked_rev(rev)?.data() |
|
34 | 34 | } |
|
35 | 35 | |
|
36 | 36 | /// Return the [`ChangelogEntry`] for the given revision number. |
|
37 | 37 | pub fn entry_for_rev( |
|
38 | 38 | &self, |
|
39 | 39 | rev: UncheckedRevision, |
|
40 | 40 | ) -> Result<ChangelogEntry, RevlogError> { |
|
41 | 41 | let revlog_entry = self.revlog.get_entry(rev)?; |
|
42 | 42 | Ok(ChangelogEntry { revlog_entry }) |
|
43 | 43 | } |
|
44 | 44 | |
|
45 | 45 | /// Same as [`Self::entry_for_rev`] for checked revisions. |
|
46 | 46 | fn entry_for_checked_rev( |
|
47 | 47 | &self, |
|
48 | 48 | rev: Revision, |
|
49 | 49 | ) -> Result<ChangelogEntry, RevlogError> { |
|
50 | 50 | let revlog_entry = self.revlog.get_entry_for_checked_rev(rev)?; |
|
51 | 51 | Ok(ChangelogEntry { revlog_entry }) |
|
52 | 52 | } |
|
53 | 53 | |
|
54 | 54 | /// Return the [`ChangelogRevisionData`] for the given revision number. |
|
55 | 55 | /// |
|
56 | 56 | /// This is a useful shortcut in case the caller does not need the |
|
57 | 57 | /// generic revlog information (parents, hashes etc). Otherwise |
|
58 | 58 | /// consider taking a [`ChangelogEntry`] with |
|
59 | 59 | /// [entry_for_rev](`Self::entry_for_rev`) and doing everything from there. |
|
60 | 60 | pub fn data_for_rev( |
|
61 | 61 | &self, |
|
62 | 62 | rev: UncheckedRevision, |
|
63 | 63 | ) -> Result<ChangelogRevisionData, RevlogError> { |
|
64 | 64 | self.entry_for_rev(rev)?.data() |
|
65 | 65 | } |
|
66 | 66 | |
|
67 | 67 | pub fn node_from_rev(&self, rev: UncheckedRevision) -> Option<&Node> { |
|
68 | 68 | self.revlog.node_from_rev(rev) |
|
69 | 69 | } |
|
70 | 70 | |
|
71 | 71 | pub fn rev_from_node( |
|
72 | 72 | &self, |
|
73 | 73 | node: NodePrefix, |
|
74 | 74 | ) -> Result<Revision, RevlogError> { |
|
75 | 75 | self.revlog.rev_from_node(node) |
|
76 | 76 | } |
|
77 | 77 | } |
|
78 | 78 | |
|
79 | impl Graph for Changelog { | |
|
80 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
|
81 | self.revlog.parents(rev) | |
|
82 | } | |
|
83 | } | |
|
84 | ||
|
79 | 85 | /// A specialized `RevlogEntry` for `changelog` data format |
|
80 | 86 | /// |
|
81 | 87 | /// This is a `RevlogEntry` with the added semantics that the associated |
|
82 | 88 | /// data should meet the requirements for `changelog`, materialized by |
|
83 | 89 | /// the fact that `data()` constructs a `ChangelogRevisionData`. |
|
84 | 90 | /// In case that promise would be broken, the `data` method returns an error. |
|
85 | 91 | #[derive(Clone)] |
|
86 | 92 | pub struct ChangelogEntry<'changelog> { |
|
87 | 93 | /// Same data, as a generic `RevlogEntry`. |
|
88 | 94 | pub(crate) revlog_entry: RevlogEntry<'changelog>, |
|
89 | 95 | } |
|
90 | 96 | |
|
91 | 97 | impl<'changelog> ChangelogEntry<'changelog> { |
|
92 | 98 | pub fn data<'a>( |
|
93 | 99 | &'a self, |
|
94 | 100 | ) -> Result<ChangelogRevisionData<'changelog>, RevlogError> { |
|
95 | 101 | let bytes = self.revlog_entry.data()?; |
|
96 | 102 | if bytes.is_empty() { |
|
97 | 103 | Ok(ChangelogRevisionData::null()) |
|
98 | 104 | } else { |
|
99 | 105 | Ok(ChangelogRevisionData::new(bytes).map_err(|err| { |
|
100 | 106 | RevlogError::Other(HgError::CorruptedRepository(format!( |
|
101 | 107 | "Invalid changelog data for revision {}: {:?}", |
|
102 | 108 | self.revlog_entry.revision(), |
|
103 | 109 | err |
|
104 | 110 | ))) |
|
105 | 111 | })?) |
|
106 | 112 | } |
|
107 | 113 | } |
|
108 | 114 | |
|
109 | 115 | /// Obtain a reference to the underlying `RevlogEntry`. |
|
110 | 116 | /// |
|
111 | 117 | /// This allows the caller to access the information that is common |
|
112 | 118 | /// to all revlog entries: revision number, node id, parent revisions etc. |
|
113 | 119 | pub fn as_revlog_entry(&self) -> &RevlogEntry { |
|
114 | 120 | &self.revlog_entry |
|
115 | 121 | } |
|
116 | 122 | |
|
117 | 123 | pub fn p1_entry(&self) -> Result<Option<ChangelogEntry>, RevlogError> { |
|
118 | 124 | Ok(self |
|
119 | 125 | .revlog_entry |
|
120 | 126 | .p1_entry()? |
|
121 | 127 | .map(|revlog_entry| Self { revlog_entry })) |
|
122 | 128 | } |
|
123 | 129 | |
|
124 | 130 | pub fn p2_entry(&self) -> Result<Option<ChangelogEntry>, RevlogError> { |
|
125 | 131 | Ok(self |
|
126 | 132 | .revlog_entry |
|
127 | 133 | .p2_entry()? |
|
128 | 134 | .map(|revlog_entry| Self { revlog_entry })) |
|
129 | 135 | } |
|
130 | 136 | } |
|
131 | 137 | |
|
132 | 138 | /// `Changelog` entry which knows how to interpret the `changelog` data bytes. |
|
133 | 139 | #[derive(PartialEq)] |
|
134 | 140 | pub struct ChangelogRevisionData<'changelog> { |
|
135 | 141 | /// The data bytes of the `changelog` entry. |
|
136 | 142 | bytes: Cow<'changelog, [u8]>, |
|
137 | 143 | /// The end offset for the hex manifest (not including the newline) |
|
138 | 144 | manifest_end: usize, |
|
139 | 145 | /// The end offset for the user+email (not including the newline) |
|
140 | 146 | user_end: usize, |
|
141 | 147 | /// The end offset for the timestamp+timezone+extras (not including the |
|
142 | 148 | /// newline) |
|
143 | 149 | timestamp_end: usize, |
|
144 | 150 | /// The end offset for the file list (not including the newline) |
|
145 | 151 | files_end: usize, |
|
146 | 152 | } |
|
147 | 153 | |
|
148 | 154 | impl<'changelog> ChangelogRevisionData<'changelog> { |
|
149 | 155 | fn new(bytes: Cow<'changelog, [u8]>) -> Result<Self, HgError> { |
|
150 | 156 | let mut line_iter = bytes.split(|b| b == &b'\n'); |
|
151 | 157 | let manifest_end = line_iter |
|
152 | 158 | .next() |
|
153 | 159 | .expect("Empty iterator from split()?") |
|
154 | 160 | .len(); |
|
155 | 161 | let user_slice = line_iter.next().ok_or_else(|| { |
|
156 | 162 | HgError::corrupted("Changeset data truncated after manifest line") |
|
157 | 163 | })?; |
|
158 | 164 | let user_end = manifest_end + 1 + user_slice.len(); |
|
159 | 165 | let timestamp_slice = line_iter.next().ok_or_else(|| { |
|
160 | 166 | HgError::corrupted("Changeset data truncated after user line") |
|
161 | 167 | })?; |
|
162 | 168 | let timestamp_end = user_end + 1 + timestamp_slice.len(); |
|
163 | 169 | let mut files_end = timestamp_end + 1; |
|
164 | 170 | loop { |
|
165 | 171 | let line = line_iter.next().ok_or_else(|| { |
|
166 | 172 | HgError::corrupted("Changeset data truncated in files list") |
|
167 | 173 | })?; |
|
168 | 174 | if line.is_empty() { |
|
169 | 175 | if files_end == bytes.len() { |
|
170 | 176 | // The list of files ended with a single newline (there |
|
171 | 177 | // should be two) |
|
172 | 178 | return Err(HgError::corrupted( |
|
173 | 179 | "Changeset data truncated after files list", |
|
174 | 180 | )); |
|
175 | 181 | } |
|
176 | 182 | files_end -= 1; |
|
177 | 183 | break; |
|
178 | 184 | } |
|
179 | 185 | files_end += line.len() + 1; |
|
180 | 186 | } |
|
181 | 187 | |
|
182 | 188 | Ok(Self { |
|
183 | 189 | bytes, |
|
184 | 190 | manifest_end, |
|
185 | 191 | user_end, |
|
186 | 192 | timestamp_end, |
|
187 | 193 | files_end, |
|
188 | 194 | }) |
|
189 | 195 | } |
|
190 | 196 | |
|
191 | 197 | fn null() -> Self { |
|
192 | 198 | Self::new(Cow::Borrowed( |
|
193 | 199 | b"0000000000000000000000000000000000000000\n\n0 0\n\n", |
|
194 | 200 | )) |
|
195 | 201 | .unwrap() |
|
196 | 202 | } |
|
197 | 203 | |
|
198 | 204 | /// Return an iterator over the lines of the entry. |
|
199 | 205 | pub fn lines(&self) -> impl Iterator<Item = &[u8]> { |
|
200 | 206 | self.bytes.split(|b| b == &b'\n') |
|
201 | 207 | } |
|
202 | 208 | |
|
203 | 209 | /// Return the node id of the `manifest` referenced by this `changelog` |
|
204 | 210 | /// entry. |
|
205 | 211 | pub fn manifest_node(&self) -> Result<Node, HgError> { |
|
206 | 212 | let manifest_node_hex = &self.bytes[..self.manifest_end]; |
|
207 | 213 | Node::from_hex_for_repo(manifest_node_hex) |
|
208 | 214 | } |
|
209 | 215 | |
|
210 | 216 | /// The full user string (usually a name followed by an email enclosed in |
|
211 | 217 | /// angle brackets) |
|
212 | 218 | pub fn user(&self) -> &[u8] { |
|
213 | 219 | &self.bytes[self.manifest_end + 1..self.user_end] |
|
214 | 220 | } |
|
215 | 221 | |
|
216 | 222 | /// The full timestamp line (timestamp in seconds, offset in seconds, and |
|
217 | 223 | /// possibly extras) |
|
218 | 224 | // TODO: We should expose this in a more useful way |
|
219 | 225 | pub fn timestamp_line(&self) -> &[u8] { |
|
220 | 226 | &self.bytes[self.user_end + 1..self.timestamp_end] |
|
221 | 227 | } |
|
222 | 228 | |
|
223 | 229 | /// The files changed in this revision. |
|
224 | 230 | pub fn files(&self) -> impl Iterator<Item = &HgPath> { |
|
225 | 231 | self.bytes[self.timestamp_end + 1..self.files_end] |
|
226 | 232 | .split(|b| b == &b'\n') |
|
227 | 233 | .map(HgPath::new) |
|
228 | 234 | } |
|
229 | 235 | |
|
230 | 236 | /// The change description. |
|
231 | 237 | pub fn description(&self) -> &[u8] { |
|
232 | 238 | &self.bytes[self.files_end + 2..] |
|
233 | 239 | } |
|
234 | 240 | } |
|
235 | 241 | |
|
236 | 242 | impl Debug for ChangelogRevisionData<'_> { |
|
237 | 243 | fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result { |
|
238 | 244 | f.debug_struct("ChangelogRevisionData") |
|
239 | 245 | .field("bytes", &debug_bytes(&self.bytes)) |
|
240 | 246 | .field("manifest", &debug_bytes(&self.bytes[..self.manifest_end])) |
|
241 | 247 | .field( |
|
242 | 248 | "user", |
|
243 | 249 | &debug_bytes( |
|
244 | 250 | &self.bytes[self.manifest_end + 1..self.user_end], |
|
245 | 251 | ), |
|
246 | 252 | ) |
|
247 | 253 | .field( |
|
248 | 254 | "timestamp", |
|
249 | 255 | &debug_bytes( |
|
250 | 256 | &self.bytes[self.user_end + 1..self.timestamp_end], |
|
251 | 257 | ), |
|
252 | 258 | ) |
|
253 | 259 | .field( |
|
254 | 260 | "files", |
|
255 | 261 | &debug_bytes( |
|
256 | 262 | &self.bytes[self.timestamp_end + 1..self.files_end], |
|
257 | 263 | ), |
|
258 | 264 | ) |
|
259 | 265 | .field( |
|
260 | 266 | "description", |
|
261 | 267 | &debug_bytes(&self.bytes[self.files_end + 2..]), |
|
262 | 268 | ) |
|
263 | 269 | .finish() |
|
264 | 270 | } |
|
265 | 271 | } |
|
266 | 272 | |
|
267 | 273 | fn debug_bytes(bytes: &[u8]) -> String { |
|
268 | 274 | String::from_utf8_lossy( |
|
269 | 275 | &bytes.iter().flat_map(|b| escape_default(*b)).collect_vec(), |
|
270 | 276 | ) |
|
271 | 277 | .to_string() |
|
272 | 278 | } |
|
273 | 279 | |
|
274 | 280 | #[cfg(test)] |
|
275 | 281 | mod tests { |
|
276 | 282 | use super::*; |
|
277 | 283 | use crate::vfs::Vfs; |
|
278 | 284 | use crate::NULL_REVISION; |
|
279 | 285 | use pretty_assertions::assert_eq; |
|
280 | 286 | |
|
281 | 287 | #[test] |
|
282 | 288 | fn test_create_changelogrevisiondata_invalid() { |
|
283 | 289 | // Completely empty |
|
284 | 290 | assert!(ChangelogRevisionData::new(Cow::Borrowed(b"abcd")).is_err()); |
|
285 | 291 | // No newline after manifest |
|
286 | 292 | assert!(ChangelogRevisionData::new(Cow::Borrowed(b"abcd")).is_err()); |
|
287 | 293 | // No newline after user |
|
288 | 294 | assert!(ChangelogRevisionData::new(Cow::Borrowed(b"abcd\n")).is_err()); |
|
289 | 295 | // No newline after timestamp |
|
290 | 296 | assert!( |
|
291 | 297 | ChangelogRevisionData::new(Cow::Borrowed(b"abcd\n\n0 0")).is_err() |
|
292 | 298 | ); |
|
293 | 299 | // Missing newline after files |
|
294 | 300 | assert!(ChangelogRevisionData::new(Cow::Borrowed( |
|
295 | 301 | b"abcd\n\n0 0\nfile1\nfile2" |
|
296 | 302 | )) |
|
297 | 303 | .is_err(),); |
|
298 | 304 | // Only one newline after files |
|
299 | 305 | assert!(ChangelogRevisionData::new(Cow::Borrowed( |
|
300 | 306 | b"abcd\n\n0 0\nfile1\nfile2\n" |
|
301 | 307 | )) |
|
302 | 308 | .is_err(),); |
|
303 | 309 | } |
|
304 | 310 | |
|
305 | 311 | #[test] |
|
306 | 312 | fn test_create_changelogrevisiondata() { |
|
307 | 313 | let data = ChangelogRevisionData::new(Cow::Borrowed( |
|
308 | 314 | b"0123456789abcdef0123456789abcdef01234567 |
|
309 | 315 | Some One <someone@example.com> |
|
310 | 316 | 0 0 |
|
311 | 317 | file1 |
|
312 | 318 | file2 |
|
313 | 319 | |
|
314 | 320 | some |
|
315 | 321 | commit |
|
316 | 322 | message", |
|
317 | 323 | )) |
|
318 | 324 | .unwrap(); |
|
319 | 325 | assert_eq!( |
|
320 | 326 | data.manifest_node().unwrap(), |
|
321 | 327 | Node::from_hex("0123456789abcdef0123456789abcdef01234567") |
|
322 | 328 | .unwrap() |
|
323 | 329 | ); |
|
324 | 330 | assert_eq!(data.user(), b"Some One <someone@example.com>"); |
|
325 | 331 | assert_eq!(data.timestamp_line(), b"0 0"); |
|
326 | 332 | assert_eq!( |
|
327 | 333 | data.files().collect_vec(), |
|
328 | 334 | vec![HgPath::new("file1"), HgPath::new("file2")] |
|
329 | 335 | ); |
|
330 | 336 | assert_eq!(data.description(), b"some\ncommit\nmessage"); |
|
331 | 337 | } |
|
332 | 338 | |
|
333 | 339 | #[test] |
|
334 | 340 | fn test_data_from_rev_null() -> Result<(), RevlogError> { |
|
335 | 341 | // an empty revlog will be enough for this case |
|
336 | 342 | let temp = tempfile::tempdir().unwrap(); |
|
337 | 343 | let vfs = Vfs { base: temp.path() }; |
|
338 | 344 | std::fs::write(temp.path().join("foo.i"), b"").unwrap(); |
|
339 | 345 | let revlog = Revlog::open(&vfs, "foo.i", None, false).unwrap(); |
|
340 | 346 | |
|
341 | 347 | let changelog = Changelog { revlog }; |
|
342 | 348 | assert_eq!( |
|
343 | 349 | changelog.data_for_rev(NULL_REVISION.into())?, |
|
344 | 350 | ChangelogRevisionData::null() |
|
345 | 351 | ); |
|
346 | 352 | // same with the intermediate entry object |
|
347 | 353 | assert_eq!( |
|
348 | 354 | changelog.entry_for_rev(NULL_REVISION.into())?.data()?, |
|
349 | 355 | ChangelogRevisionData::null() |
|
350 | 356 | ); |
|
351 | 357 | Ok(()) |
|
352 | 358 | } |
|
353 | 359 | } |
@@ -1,231 +1,239 b'' | |||
|
1 | 1 | use crate::errors::HgError; |
|
2 | 2 | use crate::exit_codes; |
|
3 | 3 | use crate::repo::Repo; |
|
4 | 4 | use crate::revlog::path_encode::path_encode; |
|
5 | 5 | use crate::revlog::NodePrefix; |
|
6 | 6 | use crate::revlog::Revision; |
|
7 | 7 | use crate::revlog::RevlogEntry; |
|
8 | 8 | use crate::revlog::{Revlog, RevlogError}; |
|
9 | 9 | use crate::utils::files::get_path_from_bytes; |
|
10 | 10 | use crate::utils::hg_path::HgPath; |
|
11 | 11 | use crate::utils::SliceExt; |
|
12 | use crate::Graph; | |
|
13 | use crate::GraphError; | |
|
12 | 14 | use crate::UncheckedRevision; |
|
13 | 15 | use std::path::PathBuf; |
|
14 | 16 | |
|
15 | 17 | /// A specialized `Revlog` to work with file data logs. |
|
16 | 18 | pub struct Filelog { |
|
17 | 19 | /// The generic `revlog` format. |
|
18 | 20 | revlog: Revlog, |
|
19 | 21 | } |
|
20 | 22 | |
|
23 | impl Graph for Filelog { | |
|
24 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
|
25 | self.revlog.parents(rev) | |
|
26 | } | |
|
27 | } | |
|
28 | ||
|
21 | 29 | impl Filelog { |
|
22 | 30 | pub fn open_vfs( |
|
23 | 31 | store_vfs: &crate::vfs::Vfs<'_>, |
|
24 | 32 | file_path: &HgPath, |
|
25 | 33 | ) -> Result<Self, HgError> { |
|
26 | 34 | let index_path = store_path(file_path, b".i"); |
|
27 | 35 | let data_path = store_path(file_path, b".d"); |
|
28 | 36 | let revlog = |
|
29 | 37 | Revlog::open(store_vfs, index_path, Some(&data_path), false)?; |
|
30 | 38 | Ok(Self { revlog }) |
|
31 | 39 | } |
|
32 | 40 | |
|
33 | 41 | pub fn open(repo: &Repo, file_path: &HgPath) -> Result<Self, HgError> { |
|
34 | 42 | Self::open_vfs(&repo.store_vfs(), file_path) |
|
35 | 43 | } |
|
36 | 44 | |
|
37 | 45 | /// The given node ID is that of the file as found in a filelog, not of a |
|
38 | 46 | /// changeset. |
|
39 | 47 | pub fn data_for_node( |
|
40 | 48 | &self, |
|
41 | 49 | file_node: impl Into<NodePrefix>, |
|
42 | 50 | ) -> Result<FilelogRevisionData, RevlogError> { |
|
43 | 51 | let file_rev = self.revlog.rev_from_node(file_node.into())?; |
|
44 | 52 | self.data_for_rev(file_rev.into()) |
|
45 | 53 | } |
|
46 | 54 | |
|
47 | 55 | /// The given revision is that of the file as found in a filelog, not of a |
|
48 | 56 | /// changeset. |
|
49 | 57 | pub fn data_for_rev( |
|
50 | 58 | &self, |
|
51 | 59 | file_rev: UncheckedRevision, |
|
52 | 60 | ) -> Result<FilelogRevisionData, RevlogError> { |
|
53 | 61 | let data: Vec<u8> = self.revlog.get_rev_data(file_rev)?.into_owned(); |
|
54 | 62 | Ok(FilelogRevisionData(data)) |
|
55 | 63 | } |
|
56 | 64 | |
|
57 | 65 | /// The given node ID is that of the file as found in a filelog, not of a |
|
58 | 66 | /// changeset. |
|
59 | 67 | pub fn entry_for_node( |
|
60 | 68 | &self, |
|
61 | 69 | file_node: impl Into<NodePrefix>, |
|
62 | 70 | ) -> Result<FilelogEntry, RevlogError> { |
|
63 | 71 | let file_rev = self.revlog.rev_from_node(file_node.into())?; |
|
64 | 72 | self.entry_for_checked_rev(file_rev) |
|
65 | 73 | } |
|
66 | 74 | |
|
67 | 75 | /// The given revision is that of the file as found in a filelog, not of a |
|
68 | 76 | /// changeset. |
|
69 | 77 | pub fn entry_for_rev( |
|
70 | 78 | &self, |
|
71 | 79 | file_rev: UncheckedRevision, |
|
72 | 80 | ) -> Result<FilelogEntry, RevlogError> { |
|
73 | 81 | Ok(FilelogEntry(self.revlog.get_entry(file_rev)?)) |
|
74 | 82 | } |
|
75 | 83 | |
|
76 | 84 | fn entry_for_checked_rev( |
|
77 | 85 | &self, |
|
78 | 86 | file_rev: Revision, |
|
79 | 87 | ) -> Result<FilelogEntry, RevlogError> { |
|
80 | 88 | Ok(FilelogEntry( |
|
81 | 89 | self.revlog.get_entry_for_checked_rev(file_rev)?, |
|
82 | 90 | )) |
|
83 | 91 | } |
|
84 | 92 | } |
|
85 | 93 | |
|
86 | 94 | fn store_path(hg_path: &HgPath, suffix: &[u8]) -> PathBuf { |
|
87 | 95 | let encoded_bytes = |
|
88 | 96 | path_encode(&[b"data/", hg_path.as_bytes(), suffix].concat()); |
|
89 | 97 | get_path_from_bytes(&encoded_bytes).into() |
|
90 | 98 | } |
|
91 | 99 | |
|
92 | 100 | pub struct FilelogEntry<'a>(RevlogEntry<'a>); |
|
93 | 101 | |
|
94 | 102 | impl FilelogEntry<'_> { |
|
95 | 103 | /// `self.data()` can be expensive, with decompression and delta |
|
96 | 104 | /// resolution. |
|
97 | 105 | /// |
|
98 | 106 | /// *Without* paying this cost, based on revlog index information |
|
99 | 107 | /// including `RevlogEntry::uncompressed_len`: |
|
100 | 108 | /// |
|
101 | 109 | /// * Returns `true` if the length that `self.data().file_data().len()` |
|
102 | 110 | /// would return is definitely **not equal** to `other_len`. |
|
103 | 111 | /// * Returns `false` if available information is inconclusive. |
|
104 | 112 | pub fn file_data_len_not_equal_to(&self, other_len: u64) -> bool { |
|
105 | 113 | // Relevant code that implement this behavior in Python code: |
|
106 | 114 | // basefilectx.cmp, filelog.size, storageutil.filerevisioncopied, |
|
107 | 115 | // revlog.size, revlog.rawsize |
|
108 | 116 | |
|
109 | 117 | // Let’s call `file_data_len` what would be returned by |
|
110 | 118 | // `self.data().file_data().len()`. |
|
111 | 119 | |
|
112 | 120 | if self.0.is_censored() { |
|
113 | 121 | let file_data_len = 0; |
|
114 | 122 | return other_len != file_data_len; |
|
115 | 123 | } |
|
116 | 124 | |
|
117 | 125 | if self.0.has_length_affecting_flag_processor() { |
|
118 | 126 | // We can’t conclude anything about `file_data_len`. |
|
119 | 127 | return false; |
|
120 | 128 | } |
|
121 | 129 | |
|
122 | 130 | // Revlog revisions (usually) have metadata for the size of |
|
123 | 131 | // their data after decompression and delta resolution |
|
124 | 132 | // as would be returned by `Revlog::get_rev_data`. |
|
125 | 133 | // |
|
126 | 134 | // For filelogs this is the file’s contents preceded by an optional |
|
127 | 135 | // metadata block. |
|
128 | 136 | let uncompressed_len = if let Some(l) = self.0.uncompressed_len() { |
|
129 | 137 | l as u64 |
|
130 | 138 | } else { |
|
131 | 139 | // The field was set to -1, the actual uncompressed len is unknown. |
|
132 | 140 | // We need to decompress to say more. |
|
133 | 141 | return false; |
|
134 | 142 | }; |
|
135 | 143 | // `uncompressed_len = file_data_len + optional_metadata_len`, |
|
136 | 144 | // so `file_data_len <= uncompressed_len`. |
|
137 | 145 | if uncompressed_len < other_len { |
|
138 | 146 | // Transitively, `file_data_len < other_len`. |
|
139 | 147 | // So `other_len != file_data_len` definitely. |
|
140 | 148 | return true; |
|
141 | 149 | } |
|
142 | 150 | |
|
143 | 151 | if uncompressed_len == other_len + 4 { |
|
144 | 152 | // It’s possible that `file_data_len == other_len` with an empty |
|
145 | 153 | // metadata block (2 start marker bytes + 2 end marker bytes). |
|
146 | 154 | // This happens when there wouldn’t otherwise be metadata, but |
|
147 | 155 | // the first 2 bytes of file data happen to match a start marker |
|
148 | 156 | // and would be ambiguous. |
|
149 | 157 | return false; |
|
150 | 158 | } |
|
151 | 159 | |
|
152 | 160 | if !self.0.has_p1() { |
|
153 | 161 | // There may or may not be copy metadata, so we can’t deduce more |
|
154 | 162 | // about `file_data_len` without computing file data. |
|
155 | 163 | return false; |
|
156 | 164 | } |
|
157 | 165 | |
|
158 | 166 | // Filelog ancestry is not meaningful in the way changelog ancestry is. |
|
159 | 167 | // It only provides hints to delta generation. |
|
160 | 168 | // p1 and p2 are set to null when making a copy or rename since |
|
161 | 169 | // contents are likely unrelatedto what might have previously existed |
|
162 | 170 | // at the destination path. |
|
163 | 171 | // |
|
164 | 172 | // Conversely, since here p1 is non-null, there is no copy metadata. |
|
165 | 173 | // Note that this reasoning may be invalidated in the presence of |
|
166 | 174 | // merges made by some previous versions of Mercurial that |
|
167 | 175 | // swapped p1 and p2. See <https://bz.mercurial-scm.org/show_bug.cgi?id=6528> |
|
168 | 176 | // and `tests/test-issue6528.t`. |
|
169 | 177 | // |
|
170 | 178 | // Since copy metadata is currently the only kind of metadata |
|
171 | 179 | // kept in revlog data of filelogs, |
|
172 | 180 | // this `FilelogEntry` does not have such metadata: |
|
173 | 181 | let file_data_len = uncompressed_len; |
|
174 | 182 | |
|
175 | 183 | file_data_len != other_len |
|
176 | 184 | } |
|
177 | 185 | |
|
178 | 186 | pub fn data(&self) -> Result<FilelogRevisionData, HgError> { |
|
179 | 187 | let data = self.0.data(); |
|
180 | 188 | match data { |
|
181 | 189 | Ok(data) => Ok(FilelogRevisionData(data.into_owned())), |
|
182 | 190 | // Errors other than `HgError` should not happen at this point |
|
183 | 191 | Err(e) => match e { |
|
184 | 192 | RevlogError::Other(hg_error) => Err(hg_error), |
|
185 | 193 | revlog_error => Err(HgError::abort( |
|
186 | 194 | revlog_error.to_string(), |
|
187 | 195 | exit_codes::ABORT, |
|
188 | 196 | None, |
|
189 | 197 | )), |
|
190 | 198 | }, |
|
191 | 199 | } |
|
192 | 200 | } |
|
193 | 201 | } |
|
194 | 202 | |
|
195 | 203 | /// The data for one revision in a filelog, uncompressed and delta-resolved. |
|
196 | 204 | pub struct FilelogRevisionData(Vec<u8>); |
|
197 | 205 | |
|
198 | 206 | impl FilelogRevisionData { |
|
199 | 207 | /// Split into metadata and data |
|
200 | 208 | pub fn split(&self) -> Result<(Option<&[u8]>, &[u8]), HgError> { |
|
201 | 209 | const DELIMITER: &[u8; 2] = &[b'\x01', b'\n']; |
|
202 | 210 | |
|
203 | 211 | if let Some(rest) = self.0.drop_prefix(DELIMITER) { |
|
204 | 212 | if let Some((metadata, data)) = rest.split_2_by_slice(DELIMITER) { |
|
205 | 213 | Ok((Some(metadata), data)) |
|
206 | 214 | } else { |
|
207 | 215 | Err(HgError::corrupted( |
|
208 | 216 | "Missing metadata end delimiter in filelog entry", |
|
209 | 217 | )) |
|
210 | 218 | } |
|
211 | 219 | } else { |
|
212 | 220 | Ok((None, &self.0)) |
|
213 | 221 | } |
|
214 | 222 | } |
|
215 | 223 | |
|
216 | 224 | /// Returns the file contents at this revision, stripped of any metadata |
|
217 | 225 | pub fn file_data(&self) -> Result<&[u8], HgError> { |
|
218 | 226 | let (_metadata, data) = self.split()?; |
|
219 | 227 | Ok(data) |
|
220 | 228 | } |
|
221 | 229 | |
|
222 | 230 | /// Consume the entry, and convert it into data, discarding any metadata, |
|
223 | 231 | /// if present. |
|
224 | 232 | pub fn into_file_data(self) -> Result<Vec<u8>, HgError> { |
|
225 | 233 | if let (Some(_metadata), data) = self.split()? { |
|
226 | 234 | Ok(data.to_owned()) |
|
227 | 235 | } else { |
|
228 | 236 | Ok(self.0) |
|
229 | 237 | } |
|
230 | 238 | } |
|
231 | 239 | } |
@@ -1,622 +1,639 b'' | |||
|
1 | 1 | use std::fmt::Debug; |
|
2 | 2 | use std::ops::Deref; |
|
3 | 3 | |
|
4 | 4 | use byteorder::{BigEndian, ByteOrder}; |
|
5 | 5 | |
|
6 | 6 | use crate::errors::HgError; |
|
7 | 7 | use crate::revlog::node::Node; |
|
8 | 8 | use crate::revlog::{Revision, NULL_REVISION}; |
|
9 | use crate::UncheckedRevision; | |
|
9 | use crate::{Graph, GraphError, RevlogIndex, UncheckedRevision}; | |
|
10 | 10 | |
|
11 | 11 | pub const INDEX_ENTRY_SIZE: usize = 64; |
|
12 | 12 | |
|
13 | 13 | pub struct IndexHeader { |
|
14 | 14 | header_bytes: [u8; 4], |
|
15 | 15 | } |
|
16 | 16 | |
|
17 | 17 | #[derive(Copy, Clone)] |
|
18 | 18 | pub struct IndexHeaderFlags { |
|
19 | 19 | flags: u16, |
|
20 | 20 | } |
|
21 | 21 | |
|
22 | 22 | /// Corresponds to the high bits of `_format_flags` in python |
|
23 | 23 | impl IndexHeaderFlags { |
|
24 | 24 | /// Corresponds to FLAG_INLINE_DATA in python |
|
25 | 25 | pub fn is_inline(self) -> bool { |
|
26 | 26 | self.flags & 1 != 0 |
|
27 | 27 | } |
|
28 | 28 | /// Corresponds to FLAG_GENERALDELTA in python |
|
29 | 29 | pub fn uses_generaldelta(self) -> bool { |
|
30 | 30 | self.flags & 2 != 0 |
|
31 | 31 | } |
|
32 | 32 | } |
|
33 | 33 | |
|
34 | 34 | /// Corresponds to the INDEX_HEADER structure, |
|
35 | 35 | /// which is parsed as a `header` variable in `_loadindex` in `revlog.py` |
|
36 | 36 | impl IndexHeader { |
|
37 | 37 | fn format_flags(&self) -> IndexHeaderFlags { |
|
38 | 38 | // No "unknown flags" check here, unlike in python. Maybe there should |
|
39 | 39 | // be. |
|
40 | 40 | IndexHeaderFlags { |
|
41 | 41 | flags: BigEndian::read_u16(&self.header_bytes[0..2]), |
|
42 | 42 | } |
|
43 | 43 | } |
|
44 | 44 | |
|
45 | 45 | /// The only revlog version currently supported by rhg. |
|
46 | 46 | const REVLOGV1: u16 = 1; |
|
47 | 47 | |
|
48 | 48 | /// Corresponds to `_format_version` in Python. |
|
49 | 49 | fn format_version(&self) -> u16 { |
|
50 | 50 | BigEndian::read_u16(&self.header_bytes[2..4]) |
|
51 | 51 | } |
|
52 | 52 | |
|
53 | 53 | const EMPTY_INDEX_HEADER: IndexHeader = IndexHeader { |
|
54 | 54 | // We treat an empty file as a valid index with no entries. |
|
55 | 55 | // Here we make an arbitrary choice of what we assume the format of the |
|
56 | 56 | // index to be (V1, using generaldelta). |
|
57 | 57 | // This doesn't matter too much, since we're only doing read-only |
|
58 | 58 | // access. but the value corresponds to the `new_header` variable in |
|
59 | 59 | // `revlog.py`, `_loadindex` |
|
60 | 60 | header_bytes: [0, 3, 0, 1], |
|
61 | 61 | }; |
|
62 | 62 | |
|
63 | 63 | fn parse(index_bytes: &[u8]) -> Result<IndexHeader, HgError> { |
|
64 | 64 | if index_bytes.is_empty() { |
|
65 | 65 | return Ok(IndexHeader::EMPTY_INDEX_HEADER); |
|
66 | 66 | } |
|
67 | 67 | if index_bytes.len() < 4 { |
|
68 | 68 | return Err(HgError::corrupted( |
|
69 | 69 | "corrupted revlog: can't read the index format header", |
|
70 | 70 | )); |
|
71 | 71 | } |
|
72 | 72 | Ok(IndexHeader { |
|
73 | 73 | header_bytes: { |
|
74 | 74 | let bytes: [u8; 4] = |
|
75 | 75 | index_bytes[0..4].try_into().expect("impossible"); |
|
76 | 76 | bytes |
|
77 | 77 | }, |
|
78 | 78 | }) |
|
79 | 79 | } |
|
80 | 80 | } |
|
81 | 81 | |
|
82 | 82 | /// A Revlog index |
|
83 | 83 | pub struct Index { |
|
84 | 84 | bytes: Box<dyn Deref<Target = [u8]> + Send>, |
|
85 | 85 | /// Offsets of starts of index blocks. |
|
86 | 86 | /// Only needed when the index is interleaved with data. |
|
87 | 87 | offsets: Option<Vec<usize>>, |
|
88 | 88 | uses_generaldelta: bool, |
|
89 | 89 | } |
|
90 | 90 | |
|
91 | 91 | impl Debug for Index { |
|
92 | 92 | fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { |
|
93 | 93 | f.debug_struct("Index") |
|
94 | 94 | .field("offsets", &self.offsets) |
|
95 | 95 | .field("uses_generaldelta", &self.uses_generaldelta) |
|
96 | 96 | .finish() |
|
97 | 97 | } |
|
98 | 98 | } |
|
99 | 99 | |
|
100 | impl Graph for Index { | |
|
101 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
|
102 | let err = || GraphError::ParentOutOfRange(rev); | |
|
103 | match self.get_entry(rev) { | |
|
104 | Some(entry) => { | |
|
105 | // The C implementation checks that the parents are valid | |
|
106 | // before returning | |
|
107 | Ok([ | |
|
108 | self.check_revision(entry.p1()).ok_or_else(err)?, | |
|
109 | self.check_revision(entry.p2()).ok_or_else(err)?, | |
|
110 | ]) | |
|
111 | } | |
|
112 | None => Ok([NULL_REVISION, NULL_REVISION]), | |
|
113 | } | |
|
114 | } | |
|
115 | } | |
|
116 | ||
|
100 | 117 | impl Index { |
|
101 | 118 | /// Create an index from bytes. |
|
102 | 119 | /// Calculate the start of each entry when is_inline is true. |
|
103 | 120 | pub fn new( |
|
104 | 121 | bytes: Box<dyn Deref<Target = [u8]> + Send>, |
|
105 | 122 | ) -> Result<Self, HgError> { |
|
106 | 123 | let header = IndexHeader::parse(bytes.as_ref())?; |
|
107 | 124 | |
|
108 | 125 | if header.format_version() != IndexHeader::REVLOGV1 { |
|
109 | 126 | // A proper new version should have had a repo/store |
|
110 | 127 | // requirement. |
|
111 | 128 | return Err(HgError::corrupted("unsupported revlog version")); |
|
112 | 129 | } |
|
113 | 130 | |
|
114 | 131 | // This is only correct because we know version is REVLOGV1. |
|
115 | 132 | // In v2 we always use generaldelta, while in v0 we never use |
|
116 | 133 | // generaldelta. Similar for [is_inline] (it's only used in v1). |
|
117 | 134 | let uses_generaldelta = header.format_flags().uses_generaldelta(); |
|
118 | 135 | |
|
119 | 136 | if header.format_flags().is_inline() { |
|
120 | 137 | let mut offset: usize = 0; |
|
121 | 138 | let mut offsets = Vec::new(); |
|
122 | 139 | |
|
123 | 140 | while offset + INDEX_ENTRY_SIZE <= bytes.len() { |
|
124 | 141 | offsets.push(offset); |
|
125 | 142 | let end = offset + INDEX_ENTRY_SIZE; |
|
126 | 143 | let entry = IndexEntry { |
|
127 | 144 | bytes: &bytes[offset..end], |
|
128 | 145 | offset_override: None, |
|
129 | 146 | }; |
|
130 | 147 | |
|
131 | 148 | offset += INDEX_ENTRY_SIZE + entry.compressed_len() as usize; |
|
132 | 149 | } |
|
133 | 150 | |
|
134 | 151 | if offset == bytes.len() { |
|
135 | 152 | Ok(Self { |
|
136 | 153 | bytes, |
|
137 | 154 | offsets: Some(offsets), |
|
138 | 155 | uses_generaldelta, |
|
139 | 156 | }) |
|
140 | 157 | } else { |
|
141 | 158 | Err(HgError::corrupted("unexpected inline revlog length")) |
|
142 | 159 | } |
|
143 | 160 | } else { |
|
144 | 161 | Ok(Self { |
|
145 | 162 | bytes, |
|
146 | 163 | offsets: None, |
|
147 | 164 | uses_generaldelta, |
|
148 | 165 | }) |
|
149 | 166 | } |
|
150 | 167 | } |
|
151 | 168 | |
|
152 | 169 | pub fn uses_generaldelta(&self) -> bool { |
|
153 | 170 | self.uses_generaldelta |
|
154 | 171 | } |
|
155 | 172 | |
|
156 | 173 | /// Value of the inline flag. |
|
157 | 174 | pub fn is_inline(&self) -> bool { |
|
158 | 175 | self.offsets.is_some() |
|
159 | 176 | } |
|
160 | 177 | |
|
161 | 178 | /// Return a slice of bytes if `revlog` is inline. Panic if not. |
|
162 | 179 | pub fn data(&self, start: usize, end: usize) -> &[u8] { |
|
163 | 180 | if !self.is_inline() { |
|
164 | 181 | panic!("tried to access data in the index of a revlog that is not inline"); |
|
165 | 182 | } |
|
166 | 183 | &self.bytes[start..end] |
|
167 | 184 | } |
|
168 | 185 | |
|
169 | 186 | /// Return number of entries of the revlog index. |
|
170 | 187 | pub fn len(&self) -> usize { |
|
171 | 188 | if let Some(offsets) = &self.offsets { |
|
172 | 189 | offsets.len() |
|
173 | 190 | } else { |
|
174 | 191 | self.bytes.len() / INDEX_ENTRY_SIZE |
|
175 | 192 | } |
|
176 | 193 | } |
|
177 | 194 | |
|
178 | 195 | /// Returns `true` if the `Index` has zero `entries`. |
|
179 | 196 | pub fn is_empty(&self) -> bool { |
|
180 | 197 | self.len() == 0 |
|
181 | 198 | } |
|
182 | 199 | |
|
183 | 200 | /// Return the index entry corresponding to the given revision if it |
|
184 | 201 | /// exists. |
|
185 | 202 | pub fn get_entry(&self, rev: Revision) -> Option<IndexEntry> { |
|
186 | 203 | if rev == NULL_REVISION { |
|
187 | 204 | return None; |
|
188 | 205 | } |
|
189 | 206 | Some(if let Some(offsets) = &self.offsets { |
|
190 | 207 | self.get_entry_inline(rev, offsets) |
|
191 | 208 | } else { |
|
192 | 209 | self.get_entry_separated(rev) |
|
193 | 210 | }) |
|
194 | 211 | } |
|
195 | 212 | |
|
196 | 213 | fn get_entry_inline( |
|
197 | 214 | &self, |
|
198 | 215 | rev: Revision, |
|
199 | 216 | offsets: &[usize], |
|
200 | 217 | ) -> IndexEntry { |
|
201 | 218 | let start = offsets[rev as usize]; |
|
202 | 219 | let end = start + INDEX_ENTRY_SIZE; |
|
203 | 220 | let bytes = &self.bytes[start..end]; |
|
204 | 221 | |
|
205 | 222 | // See IndexEntry for an explanation of this override. |
|
206 | 223 | let offset_override = Some(end); |
|
207 | 224 | |
|
208 | 225 | IndexEntry { |
|
209 | 226 | bytes, |
|
210 | 227 | offset_override, |
|
211 | 228 | } |
|
212 | 229 | } |
|
213 | 230 | |
|
214 | 231 | fn get_entry_separated(&self, rev: Revision) -> IndexEntry { |
|
215 | 232 | let start = rev as usize * INDEX_ENTRY_SIZE; |
|
216 | 233 | let end = start + INDEX_ENTRY_SIZE; |
|
217 | 234 | let bytes = &self.bytes[start..end]; |
|
218 | 235 | |
|
219 | 236 | // Override the offset of the first revision as its bytes are used |
|
220 | 237 | // for the index's metadata (saving space because it is always 0) |
|
221 | 238 | let offset_override = if rev == 0 { Some(0) } else { None }; |
|
222 | 239 | |
|
223 | 240 | IndexEntry { |
|
224 | 241 | bytes, |
|
225 | 242 | offset_override, |
|
226 | 243 | } |
|
227 | 244 | } |
|
228 | 245 | } |
|
229 | 246 | |
|
230 | 247 | impl super::RevlogIndex for Index { |
|
231 | 248 | fn len(&self) -> usize { |
|
232 | 249 | self.len() |
|
233 | 250 | } |
|
234 | 251 | |
|
235 | 252 | fn node(&self, rev: Revision) -> Option<&Node> { |
|
236 | 253 | self.get_entry(rev).map(|entry| entry.hash()) |
|
237 | 254 | } |
|
238 | 255 | } |
|
239 | 256 | |
|
240 | 257 | #[derive(Debug)] |
|
241 | 258 | pub struct IndexEntry<'a> { |
|
242 | 259 | bytes: &'a [u8], |
|
243 | 260 | /// Allows to override the offset value of the entry. |
|
244 | 261 | /// |
|
245 | 262 | /// For interleaved index and data, the offset stored in the index |
|
246 | 263 | /// corresponds to the separated data offset. |
|
247 | 264 | /// It has to be overridden with the actual offset in the interleaved |
|
248 | 265 | /// index which is just after the index block. |
|
249 | 266 | /// |
|
250 | 267 | /// For separated index and data, the offset stored in the first index |
|
251 | 268 | /// entry is mixed with the index headers. |
|
252 | 269 | /// It has to be overridden with 0. |
|
253 | 270 | offset_override: Option<usize>, |
|
254 | 271 | } |
|
255 | 272 | |
|
256 | 273 | impl<'a> IndexEntry<'a> { |
|
257 | 274 | /// Return the offset of the data. |
|
258 | 275 | pub fn offset(&self) -> usize { |
|
259 | 276 | if let Some(offset_override) = self.offset_override { |
|
260 | 277 | offset_override |
|
261 | 278 | } else { |
|
262 | 279 | let mut bytes = [0; 8]; |
|
263 | 280 | bytes[2..8].copy_from_slice(&self.bytes[0..=5]); |
|
264 | 281 | BigEndian::read_u64(&bytes[..]) as usize |
|
265 | 282 | } |
|
266 | 283 | } |
|
267 | 284 | |
|
268 | 285 | pub fn flags(&self) -> u16 { |
|
269 | 286 | BigEndian::read_u16(&self.bytes[6..=7]) |
|
270 | 287 | } |
|
271 | 288 | |
|
272 | 289 | /// Return the compressed length of the data. |
|
273 | 290 | pub fn compressed_len(&self) -> u32 { |
|
274 | 291 | BigEndian::read_u32(&self.bytes[8..=11]) |
|
275 | 292 | } |
|
276 | 293 | |
|
277 | 294 | /// Return the uncompressed length of the data. |
|
278 | 295 | pub fn uncompressed_len(&self) -> i32 { |
|
279 | 296 | BigEndian::read_i32(&self.bytes[12..=15]) |
|
280 | 297 | } |
|
281 | 298 | |
|
282 | 299 | /// Return the revision upon which the data has been derived. |
|
283 | 300 | pub fn base_revision_or_base_of_delta_chain(&self) -> UncheckedRevision { |
|
284 | 301 | // TODO Maybe return an Option when base_revision == rev? |
|
285 | 302 | // Requires to add rev to IndexEntry |
|
286 | 303 | |
|
287 | 304 | BigEndian::read_i32(&self.bytes[16..]).into() |
|
288 | 305 | } |
|
289 | 306 | |
|
290 | 307 | pub fn link_revision(&self) -> UncheckedRevision { |
|
291 | 308 | BigEndian::read_i32(&self.bytes[20..]).into() |
|
292 | 309 | } |
|
293 | 310 | |
|
294 | 311 | pub fn p1(&self) -> UncheckedRevision { |
|
295 | 312 | BigEndian::read_i32(&self.bytes[24..]).into() |
|
296 | 313 | } |
|
297 | 314 | |
|
298 | 315 | pub fn p2(&self) -> UncheckedRevision { |
|
299 | 316 | BigEndian::read_i32(&self.bytes[28..]).into() |
|
300 | 317 | } |
|
301 | 318 | |
|
302 | 319 | /// Return the hash of revision's full text. |
|
303 | 320 | /// |
|
304 | 321 | /// Currently, SHA-1 is used and only the first 20 bytes of this field |
|
305 | 322 | /// are used. |
|
306 | 323 | pub fn hash(&self) -> &'a Node { |
|
307 | 324 | (&self.bytes[32..52]).try_into().unwrap() |
|
308 | 325 | } |
|
309 | 326 | } |
|
310 | 327 | |
|
311 | 328 | #[cfg(test)] |
|
312 | 329 | mod tests { |
|
313 | 330 | use super::*; |
|
314 | 331 | use crate::node::NULL_NODE; |
|
315 | 332 | |
|
316 | 333 | #[cfg(test)] |
|
317 | 334 | #[derive(Debug, Copy, Clone)] |
|
318 | 335 | pub struct IndexEntryBuilder { |
|
319 | 336 | is_first: bool, |
|
320 | 337 | is_inline: bool, |
|
321 | 338 | is_general_delta: bool, |
|
322 | 339 | version: u16, |
|
323 | 340 | offset: usize, |
|
324 | 341 | compressed_len: usize, |
|
325 | 342 | uncompressed_len: usize, |
|
326 | 343 | base_revision_or_base_of_delta_chain: Revision, |
|
327 | 344 | link_revision: Revision, |
|
328 | 345 | p1: Revision, |
|
329 | 346 | p2: Revision, |
|
330 | 347 | node: Node, |
|
331 | 348 | } |
|
332 | 349 | |
|
333 | 350 | #[cfg(test)] |
|
334 | 351 | impl IndexEntryBuilder { |
|
335 | 352 | #[allow(clippy::new_without_default)] |
|
336 | 353 | pub fn new() -> Self { |
|
337 | 354 | Self { |
|
338 | 355 | is_first: false, |
|
339 | 356 | is_inline: false, |
|
340 | 357 | is_general_delta: true, |
|
341 | 358 | version: 1, |
|
342 | 359 | offset: 0, |
|
343 | 360 | compressed_len: 0, |
|
344 | 361 | uncompressed_len: 0, |
|
345 | 362 | base_revision_or_base_of_delta_chain: 0, |
|
346 | 363 | link_revision: 0, |
|
347 | 364 | p1: NULL_REVISION, |
|
348 | 365 | p2: NULL_REVISION, |
|
349 | 366 | node: NULL_NODE, |
|
350 | 367 | } |
|
351 | 368 | } |
|
352 | 369 | |
|
353 | 370 | pub fn is_first(&mut self, value: bool) -> &mut Self { |
|
354 | 371 | self.is_first = value; |
|
355 | 372 | self |
|
356 | 373 | } |
|
357 | 374 | |
|
358 | 375 | pub fn with_inline(&mut self, value: bool) -> &mut Self { |
|
359 | 376 | self.is_inline = value; |
|
360 | 377 | self |
|
361 | 378 | } |
|
362 | 379 | |
|
363 | 380 | pub fn with_general_delta(&mut self, value: bool) -> &mut Self { |
|
364 | 381 | self.is_general_delta = value; |
|
365 | 382 | self |
|
366 | 383 | } |
|
367 | 384 | |
|
368 | 385 | pub fn with_version(&mut self, value: u16) -> &mut Self { |
|
369 | 386 | self.version = value; |
|
370 | 387 | self |
|
371 | 388 | } |
|
372 | 389 | |
|
373 | 390 | pub fn with_offset(&mut self, value: usize) -> &mut Self { |
|
374 | 391 | self.offset = value; |
|
375 | 392 | self |
|
376 | 393 | } |
|
377 | 394 | |
|
378 | 395 | pub fn with_compressed_len(&mut self, value: usize) -> &mut Self { |
|
379 | 396 | self.compressed_len = value; |
|
380 | 397 | self |
|
381 | 398 | } |
|
382 | 399 | |
|
383 | 400 | pub fn with_uncompressed_len(&mut self, value: usize) -> &mut Self { |
|
384 | 401 | self.uncompressed_len = value; |
|
385 | 402 | self |
|
386 | 403 | } |
|
387 | 404 | |
|
388 | 405 | pub fn with_base_revision_or_base_of_delta_chain( |
|
389 | 406 | &mut self, |
|
390 | 407 | value: Revision, |
|
391 | 408 | ) -> &mut Self { |
|
392 | 409 | self.base_revision_or_base_of_delta_chain = value; |
|
393 | 410 | self |
|
394 | 411 | } |
|
395 | 412 | |
|
396 | 413 | pub fn with_link_revision(&mut self, value: Revision) -> &mut Self { |
|
397 | 414 | self.link_revision = value; |
|
398 | 415 | self |
|
399 | 416 | } |
|
400 | 417 | |
|
401 | 418 | pub fn with_p1(&mut self, value: Revision) -> &mut Self { |
|
402 | 419 | self.p1 = value; |
|
403 | 420 | self |
|
404 | 421 | } |
|
405 | 422 | |
|
406 | 423 | pub fn with_p2(&mut self, value: Revision) -> &mut Self { |
|
407 | 424 | self.p2 = value; |
|
408 | 425 | self |
|
409 | 426 | } |
|
410 | 427 | |
|
411 | 428 | pub fn with_node(&mut self, value: Node) -> &mut Self { |
|
412 | 429 | self.node = value; |
|
413 | 430 | self |
|
414 | 431 | } |
|
415 | 432 | |
|
416 | 433 | pub fn build(&self) -> Vec<u8> { |
|
417 | 434 | let mut bytes = Vec::with_capacity(INDEX_ENTRY_SIZE); |
|
418 | 435 | if self.is_first { |
|
419 | 436 | bytes.extend(&match (self.is_general_delta, self.is_inline) { |
|
420 | 437 | (false, false) => [0u8, 0], |
|
421 | 438 | (false, true) => [0u8, 1], |
|
422 | 439 | (true, false) => [0u8, 2], |
|
423 | 440 | (true, true) => [0u8, 3], |
|
424 | 441 | }); |
|
425 | 442 | bytes.extend(&self.version.to_be_bytes()); |
|
426 | 443 | // Remaining offset bytes. |
|
427 | 444 | bytes.extend(&[0u8; 2]); |
|
428 | 445 | } else { |
|
429 | 446 | // Offset stored on 48 bits (6 bytes) |
|
430 | 447 | bytes.extend(&(self.offset as u64).to_be_bytes()[2..]); |
|
431 | 448 | } |
|
432 | 449 | bytes.extend(&[0u8; 2]); // Revision flags. |
|
433 | 450 | bytes.extend(&(self.compressed_len as u32).to_be_bytes()); |
|
434 | 451 | bytes.extend(&(self.uncompressed_len as u32).to_be_bytes()); |
|
435 | 452 | bytes.extend( |
|
436 | 453 | &self.base_revision_or_base_of_delta_chain.to_be_bytes(), |
|
437 | 454 | ); |
|
438 | 455 | bytes.extend(&self.link_revision.to_be_bytes()); |
|
439 | 456 | bytes.extend(&self.p1.to_be_bytes()); |
|
440 | 457 | bytes.extend(&self.p2.to_be_bytes()); |
|
441 | 458 | bytes.extend(self.node.as_bytes()); |
|
442 | 459 | bytes.extend(vec![0u8; 12]); |
|
443 | 460 | bytes |
|
444 | 461 | } |
|
445 | 462 | } |
|
446 | 463 | |
|
447 | 464 | pub fn is_inline(index_bytes: &[u8]) -> bool { |
|
448 | 465 | IndexHeader::parse(index_bytes) |
|
449 | 466 | .expect("too short") |
|
450 | 467 | .format_flags() |
|
451 | 468 | .is_inline() |
|
452 | 469 | } |
|
453 | 470 | |
|
454 | 471 | pub fn uses_generaldelta(index_bytes: &[u8]) -> bool { |
|
455 | 472 | IndexHeader::parse(index_bytes) |
|
456 | 473 | .expect("too short") |
|
457 | 474 | .format_flags() |
|
458 | 475 | .uses_generaldelta() |
|
459 | 476 | } |
|
460 | 477 | |
|
461 | 478 | pub fn get_version(index_bytes: &[u8]) -> u16 { |
|
462 | 479 | IndexHeader::parse(index_bytes) |
|
463 | 480 | .expect("too short") |
|
464 | 481 | .format_version() |
|
465 | 482 | } |
|
466 | 483 | |
|
467 | 484 | #[test] |
|
468 | 485 | fn flags_when_no_inline_flag_test() { |
|
469 | 486 | let bytes = IndexEntryBuilder::new() |
|
470 | 487 | .is_first(true) |
|
471 | 488 | .with_general_delta(false) |
|
472 | 489 | .with_inline(false) |
|
473 | 490 | .build(); |
|
474 | 491 | |
|
475 | 492 | assert!(!is_inline(&bytes)); |
|
476 | 493 | assert!(!uses_generaldelta(&bytes)); |
|
477 | 494 | } |
|
478 | 495 | |
|
479 | 496 | #[test] |
|
480 | 497 | fn flags_when_inline_flag_test() { |
|
481 | 498 | let bytes = IndexEntryBuilder::new() |
|
482 | 499 | .is_first(true) |
|
483 | 500 | .with_general_delta(false) |
|
484 | 501 | .with_inline(true) |
|
485 | 502 | .build(); |
|
486 | 503 | |
|
487 | 504 | assert!(is_inline(&bytes)); |
|
488 | 505 | assert!(!uses_generaldelta(&bytes)); |
|
489 | 506 | } |
|
490 | 507 | |
|
491 | 508 | #[test] |
|
492 | 509 | fn flags_when_inline_and_generaldelta_flags_test() { |
|
493 | 510 | let bytes = IndexEntryBuilder::new() |
|
494 | 511 | .is_first(true) |
|
495 | 512 | .with_general_delta(true) |
|
496 | 513 | .with_inline(true) |
|
497 | 514 | .build(); |
|
498 | 515 | |
|
499 | 516 | assert!(is_inline(&bytes)); |
|
500 | 517 | assert!(uses_generaldelta(&bytes)); |
|
501 | 518 | } |
|
502 | 519 | |
|
503 | 520 | #[test] |
|
504 | 521 | fn test_offset() { |
|
505 | 522 | let bytes = IndexEntryBuilder::new().with_offset(1).build(); |
|
506 | 523 | let entry = IndexEntry { |
|
507 | 524 | bytes: &bytes, |
|
508 | 525 | offset_override: None, |
|
509 | 526 | }; |
|
510 | 527 | |
|
511 | 528 | assert_eq!(entry.offset(), 1) |
|
512 | 529 | } |
|
513 | 530 | |
|
514 | 531 | #[test] |
|
515 | 532 | fn test_with_overridden_offset() { |
|
516 | 533 | let bytes = IndexEntryBuilder::new().with_offset(1).build(); |
|
517 | 534 | let entry = IndexEntry { |
|
518 | 535 | bytes: &bytes, |
|
519 | 536 | offset_override: Some(2), |
|
520 | 537 | }; |
|
521 | 538 | |
|
522 | 539 | assert_eq!(entry.offset(), 2) |
|
523 | 540 | } |
|
524 | 541 | |
|
525 | 542 | #[test] |
|
526 | 543 | fn test_compressed_len() { |
|
527 | 544 | let bytes = IndexEntryBuilder::new().with_compressed_len(1).build(); |
|
528 | 545 | let entry = IndexEntry { |
|
529 | 546 | bytes: &bytes, |
|
530 | 547 | offset_override: None, |
|
531 | 548 | }; |
|
532 | 549 | |
|
533 | 550 | assert_eq!(entry.compressed_len(), 1) |
|
534 | 551 | } |
|
535 | 552 | |
|
536 | 553 | #[test] |
|
537 | 554 | fn test_uncompressed_len() { |
|
538 | 555 | let bytes = IndexEntryBuilder::new().with_uncompressed_len(1).build(); |
|
539 | 556 | let entry = IndexEntry { |
|
540 | 557 | bytes: &bytes, |
|
541 | 558 | offset_override: None, |
|
542 | 559 | }; |
|
543 | 560 | |
|
544 | 561 | assert_eq!(entry.uncompressed_len(), 1) |
|
545 | 562 | } |
|
546 | 563 | |
|
547 | 564 | #[test] |
|
548 | 565 | fn test_base_revision_or_base_of_delta_chain() { |
|
549 | 566 | let bytes = IndexEntryBuilder::new() |
|
550 | 567 | .with_base_revision_or_base_of_delta_chain(1) |
|
551 | 568 | .build(); |
|
552 | 569 | let entry = IndexEntry { |
|
553 | 570 | bytes: &bytes, |
|
554 | 571 | offset_override: None, |
|
555 | 572 | }; |
|
556 | 573 | |
|
557 | 574 | assert_eq!(entry.base_revision_or_base_of_delta_chain(), 1.into()) |
|
558 | 575 | } |
|
559 | 576 | |
|
560 | 577 | #[test] |
|
561 | 578 | fn link_revision_test() { |
|
562 | 579 | let bytes = IndexEntryBuilder::new().with_link_revision(123).build(); |
|
563 | 580 | |
|
564 | 581 | let entry = IndexEntry { |
|
565 | 582 | bytes: &bytes, |
|
566 | 583 | offset_override: None, |
|
567 | 584 | }; |
|
568 | 585 | |
|
569 | 586 | assert_eq!(entry.link_revision(), 123.into()); |
|
570 | 587 | } |
|
571 | 588 | |
|
572 | 589 | #[test] |
|
573 | 590 | fn p1_test() { |
|
574 | 591 | let bytes = IndexEntryBuilder::new().with_p1(123).build(); |
|
575 | 592 | |
|
576 | 593 | let entry = IndexEntry { |
|
577 | 594 | bytes: &bytes, |
|
578 | 595 | offset_override: None, |
|
579 | 596 | }; |
|
580 | 597 | |
|
581 | 598 | assert_eq!(entry.p1(), 123.into()); |
|
582 | 599 | } |
|
583 | 600 | |
|
584 | 601 | #[test] |
|
585 | 602 | fn p2_test() { |
|
586 | 603 | let bytes = IndexEntryBuilder::new().with_p2(123).build(); |
|
587 | 604 | |
|
588 | 605 | let entry = IndexEntry { |
|
589 | 606 | bytes: &bytes, |
|
590 | 607 | offset_override: None, |
|
591 | 608 | }; |
|
592 | 609 | |
|
593 | 610 | assert_eq!(entry.p2(), 123.into()); |
|
594 | 611 | } |
|
595 | 612 | |
|
596 | 613 | #[test] |
|
597 | 614 | fn node_test() { |
|
598 | 615 | let node = Node::from_hex("0123456789012345678901234567890123456789") |
|
599 | 616 | .unwrap(); |
|
600 | 617 | let bytes = IndexEntryBuilder::new().with_node(node).build(); |
|
601 | 618 | |
|
602 | 619 | let entry = IndexEntry { |
|
603 | 620 | bytes: &bytes, |
|
604 | 621 | offset_override: None, |
|
605 | 622 | }; |
|
606 | 623 | |
|
607 | 624 | assert_eq!(*entry.hash(), node); |
|
608 | 625 | } |
|
609 | 626 | |
|
610 | 627 | #[test] |
|
611 | 628 | fn version_test() { |
|
612 | 629 | let bytes = IndexEntryBuilder::new() |
|
613 | 630 | .is_first(true) |
|
614 | 631 | .with_version(2) |
|
615 | 632 | .build(); |
|
616 | 633 | |
|
617 | 634 | assert_eq!(get_version(&bytes), 2) |
|
618 | 635 | } |
|
619 | 636 | } |
|
620 | 637 | |
|
621 | 638 | #[cfg(test)] |
|
622 | 639 | pub use tests::IndexEntryBuilder; |
@@ -1,203 +1,209 b'' | |||
|
1 | 1 | use crate::errors::HgError; |
|
2 | 2 | use crate::revlog::{Node, NodePrefix}; |
|
3 | 3 | use crate::revlog::{Revlog, RevlogError}; |
|
4 | 4 | use crate::utils::hg_path::HgPath; |
|
5 | 5 | use crate::utils::SliceExt; |
|
6 | 6 | use crate::vfs::Vfs; |
|
7 | use crate::{Revision, UncheckedRevision}; | |
|
7 | use crate::{Graph, GraphError, Revision, UncheckedRevision}; | |
|
8 | 8 | |
|
9 | 9 | /// A specialized `Revlog` to work with `manifest` data format. |
|
10 | 10 | pub struct Manifestlog { |
|
11 | 11 | /// The generic `revlog` format. |
|
12 | 12 | revlog: Revlog, |
|
13 | 13 | } |
|
14 | 14 | |
|
15 | impl Graph for Manifestlog { | |
|
16 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
|
17 | self.revlog.parents(rev) | |
|
18 | } | |
|
19 | } | |
|
20 | ||
|
15 | 21 | impl Manifestlog { |
|
16 | 22 | /// Open the `manifest` of a repository given by its root. |
|
17 | 23 | pub fn open(store_vfs: &Vfs, use_nodemap: bool) -> Result<Self, HgError> { |
|
18 | 24 | let revlog = |
|
19 | 25 | Revlog::open(store_vfs, "00manifest.i", None, use_nodemap)?; |
|
20 | 26 | Ok(Self { revlog }) |
|
21 | 27 | } |
|
22 | 28 | |
|
23 | 29 | /// Return the `Manifest` for the given node ID. |
|
24 | 30 | /// |
|
25 | 31 | /// Note: this is a node ID in the manifestlog, typically found through |
|
26 | 32 | /// `ChangelogEntry::manifest_node`. It is *not* the node ID of any |
|
27 | 33 | /// changeset. |
|
28 | 34 | /// |
|
29 | 35 | /// See also `Repo::manifest_for_node` |
|
30 | 36 | pub fn data_for_node( |
|
31 | 37 | &self, |
|
32 | 38 | node: NodePrefix, |
|
33 | 39 | ) -> Result<Manifest, RevlogError> { |
|
34 | 40 | let rev = self.revlog.rev_from_node(node)?; |
|
35 | 41 | self.data_for_checked_rev(rev) |
|
36 | 42 | } |
|
37 | 43 | |
|
38 | 44 | /// Return the `Manifest` of a given revision number. |
|
39 | 45 | /// |
|
40 | 46 | /// Note: this is a revision number in the manifestlog, *not* of any |
|
41 | 47 | /// changeset. |
|
42 | 48 | /// |
|
43 | 49 | /// See also `Repo::manifest_for_rev` |
|
44 | 50 | pub fn data_for_rev( |
|
45 | 51 | &self, |
|
46 | 52 | rev: UncheckedRevision, |
|
47 | 53 | ) -> Result<Manifest, RevlogError> { |
|
48 | 54 | let bytes = self.revlog.get_rev_data(rev)?.into_owned(); |
|
49 | 55 | Ok(Manifest { bytes }) |
|
50 | 56 | } |
|
51 | 57 | |
|
52 | 58 | pub fn data_for_checked_rev( |
|
53 | 59 | &self, |
|
54 | 60 | rev: Revision, |
|
55 | 61 | ) -> Result<Manifest, RevlogError> { |
|
56 | 62 | let bytes = |
|
57 | 63 | self.revlog.get_rev_data_for_checked_rev(rev)?.into_owned(); |
|
58 | 64 | Ok(Manifest { bytes }) |
|
59 | 65 | } |
|
60 | 66 | } |
|
61 | 67 | |
|
62 | 68 | /// `Manifestlog` entry which knows how to interpret the `manifest` data bytes. |
|
63 | 69 | #[derive(Debug)] |
|
64 | 70 | pub struct Manifest { |
|
65 | 71 | /// Format for a manifest: flat sequence of variable-size entries, |
|
66 | 72 | /// sorted by path, each as: |
|
67 | 73 | /// |
|
68 | 74 | /// ```text |
|
69 | 75 | /// <path> \0 <hex_node_id> <flags> \n |
|
70 | 76 | /// ``` |
|
71 | 77 | /// |
|
72 | 78 | /// The last entry is also terminated by a newline character. |
|
73 | 79 | /// Flags is one of `b""` (the empty string), `b"x"`, `b"l"`, or `b"t"`. |
|
74 | 80 | bytes: Vec<u8>, |
|
75 | 81 | } |
|
76 | 82 | |
|
77 | 83 | impl Manifest { |
|
78 | 84 | pub fn iter( |
|
79 | 85 | &self, |
|
80 | 86 | ) -> impl Iterator<Item = Result<ManifestEntry, HgError>> { |
|
81 | 87 | self.bytes |
|
82 | 88 | .split(|b| b == &b'\n') |
|
83 | 89 | .filter(|line| !line.is_empty()) |
|
84 | 90 | .map(ManifestEntry::from_raw) |
|
85 | 91 | } |
|
86 | 92 | |
|
87 | 93 | /// If the given path is in this manifest, return its filelog node ID |
|
88 | 94 | pub fn find_by_path( |
|
89 | 95 | &self, |
|
90 | 96 | path: &HgPath, |
|
91 | 97 | ) -> Result<Option<ManifestEntry>, HgError> { |
|
92 | 98 | use std::cmp::Ordering::*; |
|
93 | 99 | let path = path.as_bytes(); |
|
94 | 100 | // Both boundaries of this `&[u8]` slice are always at the boundary of |
|
95 | 101 | // an entry |
|
96 | 102 | let mut bytes = &*self.bytes; |
|
97 | 103 | |
|
98 | 104 | // Binary search algorithm derived from `[T]::binary_search_by` |
|
99 | 105 | // <https://github.com/rust-lang/rust/blob/1.57.0/library/core/src/slice/mod.rs#L2221> |
|
100 | 106 | // except we don’t have a slice of entries. Instead we jump to the |
|
101 | 107 | // middle of the byte slice and look around for entry delimiters |
|
102 | 108 | // (newlines). |
|
103 | 109 | while let Some(entry_range) = Self::find_entry_near_middle_of(bytes)? { |
|
104 | 110 | let (entry_path, rest) = |
|
105 | 111 | ManifestEntry::split_path(&bytes[entry_range.clone()])?; |
|
106 | 112 | let cmp = entry_path.cmp(path); |
|
107 | 113 | if cmp == Less { |
|
108 | 114 | let after_newline = entry_range.end + 1; |
|
109 | 115 | bytes = &bytes[after_newline..]; |
|
110 | 116 | } else if cmp == Greater { |
|
111 | 117 | bytes = &bytes[..entry_range.start]; |
|
112 | 118 | } else { |
|
113 | 119 | return Ok(Some(ManifestEntry::from_path_and_rest( |
|
114 | 120 | entry_path, rest, |
|
115 | 121 | ))); |
|
116 | 122 | } |
|
117 | 123 | } |
|
118 | 124 | Ok(None) |
|
119 | 125 | } |
|
120 | 126 | |
|
121 | 127 | /// If there is at least one, return the byte range of an entry *excluding* |
|
122 | 128 | /// the final newline. |
|
123 | 129 | fn find_entry_near_middle_of( |
|
124 | 130 | bytes: &[u8], |
|
125 | 131 | ) -> Result<Option<std::ops::Range<usize>>, HgError> { |
|
126 | 132 | let len = bytes.len(); |
|
127 | 133 | if len > 0 { |
|
128 | 134 | let middle = bytes.len() / 2; |
|
129 | 135 | // Integer division rounds down, so `middle < len`. |
|
130 | 136 | let (before, after) = bytes.split_at(middle); |
|
131 | 137 | let is_newline = |&byte: &u8| byte == b'\n'; |
|
132 | 138 | let entry_start = match before.iter().rposition(is_newline) { |
|
133 | 139 | Some(i) => i + 1, |
|
134 | 140 | None => 0, // We choose the first entry in `bytes` |
|
135 | 141 | }; |
|
136 | 142 | let entry_end = match after.iter().position(is_newline) { |
|
137 | 143 | Some(i) => { |
|
138 | 144 | // No `+ 1` here to exclude this newline from the range |
|
139 | 145 | middle + i |
|
140 | 146 | } |
|
141 | 147 | None => { |
|
142 | 148 | // In a well-formed manifest: |
|
143 | 149 | // |
|
144 | 150 | // * Since `len > 0`, `bytes` contains at least one entry |
|
145 | 151 | // * Every entry ends with a newline |
|
146 | 152 | // * Since `middle < len`, `after` contains at least the |
|
147 | 153 | // newline at the end of the last entry of `bytes`. |
|
148 | 154 | // |
|
149 | 155 | // We didn’t find a newline, so this manifest is not |
|
150 | 156 | // well-formed. |
|
151 | 157 | return Err(HgError::corrupted( |
|
152 | 158 | "manifest entry without \\n delimiter", |
|
153 | 159 | )); |
|
154 | 160 | } |
|
155 | 161 | }; |
|
156 | 162 | Ok(Some(entry_start..entry_end)) |
|
157 | 163 | } else { |
|
158 | 164 | // len == 0 |
|
159 | 165 | Ok(None) |
|
160 | 166 | } |
|
161 | 167 | } |
|
162 | 168 | } |
|
163 | 169 | |
|
164 | 170 | /// `Manifestlog` entry which knows how to interpret the `manifest` data bytes. |
|
165 | 171 | #[derive(Debug)] |
|
166 | 172 | pub struct ManifestEntry<'manifest> { |
|
167 | 173 | pub path: &'manifest HgPath, |
|
168 | 174 | pub hex_node_id: &'manifest [u8], |
|
169 | 175 | |
|
170 | 176 | /// `Some` values are b'x', b'l', or 't' |
|
171 | 177 | pub flags: Option<u8>, |
|
172 | 178 | } |
|
173 | 179 | |
|
174 | 180 | impl<'a> ManifestEntry<'a> { |
|
175 | 181 | fn split_path(bytes: &[u8]) -> Result<(&[u8], &[u8]), HgError> { |
|
176 | 182 | bytes.split_2(b'\0').ok_or_else(|| { |
|
177 | 183 | HgError::corrupted("manifest entry without \\0 delimiter") |
|
178 | 184 | }) |
|
179 | 185 | } |
|
180 | 186 | |
|
181 | 187 | fn from_path_and_rest(path: &'a [u8], rest: &'a [u8]) -> Self { |
|
182 | 188 | let (hex_node_id, flags) = match rest.split_last() { |
|
183 | 189 | Some((&b'x', rest)) => (rest, Some(b'x')), |
|
184 | 190 | Some((&b'l', rest)) => (rest, Some(b'l')), |
|
185 | 191 | Some((&b't', rest)) => (rest, Some(b't')), |
|
186 | 192 | _ => (rest, None), |
|
187 | 193 | }; |
|
188 | 194 | Self { |
|
189 | 195 | path: HgPath::new(path), |
|
190 | 196 | hex_node_id, |
|
191 | 197 | flags, |
|
192 | 198 | } |
|
193 | 199 | } |
|
194 | 200 | |
|
195 | 201 | fn from_raw(bytes: &'a [u8]) -> Result<Self, HgError> { |
|
196 | 202 | let (path, rest) = Self::split_path(bytes)?; |
|
197 | 203 | Ok(Self::from_path_and_rest(path, rest)) |
|
198 | 204 | } |
|
199 | 205 | |
|
200 | 206 | pub fn node_id(&self) -> Result<Node, HgError> { |
|
201 | 207 | Node::from_hex_for_repo(self.hex_node_id) |
|
202 | 208 | } |
|
203 | 209 | } |
@@ -1,904 +1,910 b'' | |||
|
1 | 1 | // Copyright 2018-2023 Georges Racinet <georges.racinet@octobus.net> |
|
2 | 2 | // and Mercurial contributors |
|
3 | 3 | // |
|
4 | 4 | // This software may be used and distributed according to the terms of the |
|
5 | 5 | // GNU General Public License version 2 or any later version. |
|
6 | 6 | //! Mercurial concepts for handling revision history |
|
7 | 7 | |
|
8 | 8 | pub mod node; |
|
9 | 9 | pub mod nodemap; |
|
10 | 10 | mod nodemap_docket; |
|
11 | 11 | pub mod path_encode; |
|
12 | 12 | pub use node::{FromHexError, Node, NodePrefix}; |
|
13 | 13 | pub mod changelog; |
|
14 | 14 | pub mod filelog; |
|
15 | 15 | pub mod index; |
|
16 | 16 | pub mod manifest; |
|
17 | 17 | pub mod patch; |
|
18 | 18 | |
|
19 | 19 | use std::borrow::Cow; |
|
20 | 20 | use std::io::Read; |
|
21 | 21 | use std::ops::Deref; |
|
22 | 22 | use std::path::Path; |
|
23 | 23 | |
|
24 | 24 | use flate2::read::ZlibDecoder; |
|
25 | 25 | use sha1::{Digest, Sha1}; |
|
26 | 26 | use std::cell::RefCell; |
|
27 | 27 | use zstd; |
|
28 | 28 | |
|
29 | 29 | use self::node::{NODE_BYTES_LENGTH, NULL_NODE}; |
|
30 | 30 | use self::nodemap_docket::NodeMapDocket; |
|
31 | 31 | use super::index::Index; |
|
32 | 32 | use super::nodemap::{NodeMap, NodeMapError}; |
|
33 | 33 | use crate::errors::HgError; |
|
34 | 34 | use crate::vfs::Vfs; |
|
35 | 35 | |
|
36 | 36 | /// Mercurial revision numbers |
|
37 | 37 | /// |
|
38 | 38 | /// As noted in revlog.c, revision numbers are actually encoded in |
|
39 | 39 | /// 4 bytes, and are liberally converted to ints, whence the i32 |
|
40 | 40 | pub type Revision = i32; |
|
41 | 41 | |
|
42 | 42 | /// Unchecked Mercurial revision numbers. |
|
43 | 43 | /// |
|
44 | 44 | /// Values of this type have no guarantee of being a valid revision number |
|
45 | 45 | /// in any context. Use method `check_revision` to get a valid revision within |
|
46 | 46 | /// the appropriate index object. |
|
47 | 47 | /// |
|
48 | 48 | /// As noted in revlog.c, revision numbers are actually encoded in |
|
49 | 49 | /// 4 bytes, and are liberally converted to ints, whence the i32 |
|
50 | 50 | #[derive( |
|
51 | 51 | Debug, |
|
52 | 52 | derive_more::Display, |
|
53 | 53 | Clone, |
|
54 | 54 | Copy, |
|
55 | 55 | Hash, |
|
56 | 56 | PartialEq, |
|
57 | 57 | Eq, |
|
58 | 58 | PartialOrd, |
|
59 | 59 | Ord, |
|
60 | 60 | )] |
|
61 | 61 | pub struct UncheckedRevision(i32); |
|
62 | 62 | |
|
63 | 63 | impl From<Revision> for UncheckedRevision { |
|
64 | 64 | fn from(value: Revision) -> Self { |
|
65 | 65 | Self(value) |
|
66 | 66 | } |
|
67 | 67 | } |
|
68 | 68 | |
|
69 | 69 | /// Marker expressing the absence of a parent |
|
70 | 70 | /// |
|
71 | 71 | /// Independently of the actual representation, `NULL_REVISION` is guaranteed |
|
72 | 72 | /// to be smaller than all existing revisions. |
|
73 | 73 | pub const NULL_REVISION: Revision = -1; |
|
74 | 74 | |
|
75 | 75 | /// Same as `mercurial.node.wdirrev` |
|
76 | 76 | /// |
|
77 | 77 | /// This is also equal to `i32::max_value()`, but it's better to spell |
|
78 | 78 | /// it out explicitely, same as in `mercurial.node` |
|
79 | 79 | #[allow(clippy::unreadable_literal)] |
|
80 | 80 | pub const WORKING_DIRECTORY_REVISION: UncheckedRevision = |
|
81 | 81 | UncheckedRevision(0x7fffffff); |
|
82 | 82 | |
|
83 | 83 | pub const WORKING_DIRECTORY_HEX: &str = |
|
84 | 84 | "ffffffffffffffffffffffffffffffffffffffff"; |
|
85 | 85 | |
|
86 | 86 | /// The simplest expression of what we need of Mercurial DAGs. |
|
87 | 87 | pub trait Graph { |
|
88 | 88 | /// Return the two parents of the given `Revision`. |
|
89 | 89 | /// |
|
90 | 90 | /// Each of the parents can be independently `NULL_REVISION` |
|
91 | 91 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError>; |
|
92 | 92 | } |
|
93 | 93 | |
|
94 | 94 | #[derive(Clone, Debug, PartialEq)] |
|
95 | 95 | pub enum GraphError { |
|
96 | 96 | ParentOutOfRange(Revision), |
|
97 | 97 | } |
|
98 | 98 | |
|
99 | 99 | /// The Mercurial Revlog Index |
|
100 | 100 | /// |
|
101 | 101 | /// This is currently limited to the minimal interface that is needed for |
|
102 | 102 | /// the [`nodemap`](nodemap/index.html) module |
|
103 | 103 | pub trait RevlogIndex { |
|
104 | 104 | /// Total number of Revisions referenced in this index |
|
105 | 105 | fn len(&self) -> usize; |
|
106 | 106 | |
|
107 | 107 | fn is_empty(&self) -> bool { |
|
108 | 108 | self.len() == 0 |
|
109 | 109 | } |
|
110 | 110 | |
|
111 | 111 | /// Return a reference to the Node or `None` for `NULL_REVISION` |
|
112 | 112 | fn node(&self, rev: Revision) -> Option<&Node>; |
|
113 | 113 | |
|
114 | 114 | /// Return a [`Revision`] if `rev` is a valid revision number for this |
|
115 | 115 | /// index |
|
116 | 116 | fn check_revision(&self, rev: UncheckedRevision) -> Option<Revision> { |
|
117 | 117 | let rev = rev.0; |
|
118 | 118 | |
|
119 | 119 | if rev == NULL_REVISION || (rev >= 0 && (rev as usize) < self.len()) { |
|
120 | 120 | Some(rev) |
|
121 | 121 | } else { |
|
122 | 122 | None |
|
123 | 123 | } |
|
124 | 124 | } |
|
125 | 125 | } |
|
126 | 126 | |
|
127 | 127 | const REVISION_FLAG_CENSORED: u16 = 1 << 15; |
|
128 | 128 | const REVISION_FLAG_ELLIPSIS: u16 = 1 << 14; |
|
129 | 129 | const REVISION_FLAG_EXTSTORED: u16 = 1 << 13; |
|
130 | 130 | const REVISION_FLAG_HASCOPIESINFO: u16 = 1 << 12; |
|
131 | 131 | |
|
132 | 132 | // Keep this in sync with REVIDX_KNOWN_FLAGS in |
|
133 | 133 | // mercurial/revlogutils/flagutil.py |
|
134 | 134 | const REVIDX_KNOWN_FLAGS: u16 = REVISION_FLAG_CENSORED |
|
135 | 135 | | REVISION_FLAG_ELLIPSIS |
|
136 | 136 | | REVISION_FLAG_EXTSTORED |
|
137 | 137 | | REVISION_FLAG_HASCOPIESINFO; |
|
138 | 138 | |
|
139 | 139 | const NULL_REVLOG_ENTRY_FLAGS: u16 = 0; |
|
140 | 140 | |
|
141 | 141 | #[derive(Debug, derive_more::From, derive_more::Display)] |
|
142 | 142 | pub enum RevlogError { |
|
143 | 143 | InvalidRevision, |
|
144 | 144 | /// Working directory is not supported |
|
145 | 145 | WDirUnsupported, |
|
146 | 146 | /// Found more than one entry whose ID match the requested prefix |
|
147 | 147 | AmbiguousPrefix, |
|
148 | 148 | #[from] |
|
149 | 149 | Other(HgError), |
|
150 | 150 | } |
|
151 | 151 | |
|
152 | 152 | impl From<NodeMapError> for RevlogError { |
|
153 | 153 | fn from(error: NodeMapError) -> Self { |
|
154 | 154 | match error { |
|
155 | 155 | NodeMapError::MultipleResults => RevlogError::AmbiguousPrefix, |
|
156 | 156 | NodeMapError::RevisionNotInIndex(rev) => RevlogError::corrupted( |
|
157 | 157 | format!("nodemap point to revision {} not in index", rev), |
|
158 | 158 | ), |
|
159 | 159 | } |
|
160 | 160 | } |
|
161 | 161 | } |
|
162 | 162 | |
|
163 | 163 | fn corrupted<S: AsRef<str>>(context: S) -> HgError { |
|
164 | 164 | HgError::corrupted(format!("corrupted revlog, {}", context.as_ref())) |
|
165 | 165 | } |
|
166 | 166 | |
|
167 | 167 | impl RevlogError { |
|
168 | 168 | fn corrupted<S: AsRef<str>>(context: S) -> Self { |
|
169 | 169 | RevlogError::Other(corrupted(context)) |
|
170 | 170 | } |
|
171 | 171 | } |
|
172 | 172 | |
|
173 | 173 | /// Read only implementation of revlog. |
|
174 | 174 | pub struct Revlog { |
|
175 | 175 | /// When index and data are not interleaved: bytes of the revlog index. |
|
176 | 176 | /// When index and data are interleaved: bytes of the revlog index and |
|
177 | 177 | /// data. |
|
178 | 178 | index: Index, |
|
179 | 179 | /// When index and data are not interleaved: bytes of the revlog data |
|
180 | 180 | data_bytes: Option<Box<dyn Deref<Target = [u8]> + Send>>, |
|
181 | 181 | /// When present on disk: the persistent nodemap for this revlog |
|
182 | 182 | nodemap: Option<nodemap::NodeTree>, |
|
183 | 183 | } |
|
184 | 184 | |
|
185 | impl Graph for Revlog { | |
|
186 | fn parents(&self, rev: Revision) -> Result<[Revision; 2], GraphError> { | |
|
187 | self.index.parents(rev) | |
|
188 | } | |
|
189 | } | |
|
190 | ||
|
185 | 191 | impl Revlog { |
|
186 | 192 | /// Open a revlog index file. |
|
187 | 193 | /// |
|
188 | 194 | /// It will also open the associated data file if index and data are not |
|
189 | 195 | /// interleaved. |
|
190 | 196 | pub fn open( |
|
191 | 197 | store_vfs: &Vfs, |
|
192 | 198 | index_path: impl AsRef<Path>, |
|
193 | 199 | data_path: Option<&Path>, |
|
194 | 200 | use_nodemap: bool, |
|
195 | 201 | ) -> Result<Self, HgError> { |
|
196 | 202 | let index_path = index_path.as_ref(); |
|
197 | 203 | let index = { |
|
198 | 204 | match store_vfs.mmap_open_opt(&index_path)? { |
|
199 | 205 | None => Index::new(Box::new(vec![])), |
|
200 | 206 | Some(index_mmap) => { |
|
201 | 207 | let index = Index::new(Box::new(index_mmap))?; |
|
202 | 208 | Ok(index) |
|
203 | 209 | } |
|
204 | 210 | } |
|
205 | 211 | }?; |
|
206 | 212 | |
|
207 | 213 | let default_data_path = index_path.with_extension("d"); |
|
208 | 214 | |
|
209 | 215 | // type annotation required |
|
210 | 216 | // won't recognize Mmap as Deref<Target = [u8]> |
|
211 | 217 | let data_bytes: Option<Box<dyn Deref<Target = [u8]> + Send>> = |
|
212 | 218 | if index.is_inline() { |
|
213 | 219 | None |
|
214 | 220 | } else { |
|
215 | 221 | let data_path = data_path.unwrap_or(&default_data_path); |
|
216 | 222 | let data_mmap = store_vfs.mmap_open(data_path)?; |
|
217 | 223 | Some(Box::new(data_mmap)) |
|
218 | 224 | }; |
|
219 | 225 | |
|
220 | 226 | let nodemap = if index.is_inline() || !use_nodemap { |
|
221 | 227 | None |
|
222 | 228 | } else { |
|
223 | 229 | NodeMapDocket::read_from_file(store_vfs, index_path)?.map( |
|
224 | 230 | |(docket, data)| { |
|
225 | 231 | nodemap::NodeTree::load_bytes( |
|
226 | 232 | Box::new(data), |
|
227 | 233 | docket.data_length, |
|
228 | 234 | ) |
|
229 | 235 | }, |
|
230 | 236 | ) |
|
231 | 237 | }; |
|
232 | 238 | |
|
233 | 239 | Ok(Revlog { |
|
234 | 240 | index, |
|
235 | 241 | data_bytes, |
|
236 | 242 | nodemap, |
|
237 | 243 | }) |
|
238 | 244 | } |
|
239 | 245 | |
|
240 | 246 | /// Return number of entries of the `Revlog`. |
|
241 | 247 | pub fn len(&self) -> usize { |
|
242 | 248 | self.index.len() |
|
243 | 249 | } |
|
244 | 250 | |
|
245 | 251 | /// Returns `true` if the `Revlog` has zero `entries`. |
|
246 | 252 | pub fn is_empty(&self) -> bool { |
|
247 | 253 | self.index.is_empty() |
|
248 | 254 | } |
|
249 | 255 | |
|
250 | 256 | /// Returns the node ID for the given revision number, if it exists in this |
|
251 | 257 | /// revlog |
|
252 | 258 | pub fn node_from_rev(&self, rev: UncheckedRevision) -> Option<&Node> { |
|
253 | 259 | if rev == NULL_REVISION.into() { |
|
254 | 260 | return Some(&NULL_NODE); |
|
255 | 261 | } |
|
256 | 262 | let rev = self.index.check_revision(rev)?; |
|
257 | 263 | Some(self.index.get_entry(rev)?.hash()) |
|
258 | 264 | } |
|
259 | 265 | |
|
260 | 266 | /// Return the revision number for the given node ID, if it exists in this |
|
261 | 267 | /// revlog |
|
262 | 268 | pub fn rev_from_node( |
|
263 | 269 | &self, |
|
264 | 270 | node: NodePrefix, |
|
265 | 271 | ) -> Result<Revision, RevlogError> { |
|
266 | 272 | let looked_up = if let Some(nodemap) = &self.nodemap { |
|
267 | 273 | nodemap |
|
268 | 274 | .find_bin(&self.index, node)? |
|
269 | 275 | .ok_or(RevlogError::InvalidRevision) |
|
270 | 276 | } else { |
|
271 | 277 | self.rev_from_node_no_persistent_nodemap(node) |
|
272 | 278 | }; |
|
273 | 279 | |
|
274 | 280 | if node.is_prefix_of(&NULL_NODE) { |
|
275 | 281 | return match looked_up { |
|
276 | 282 | Ok(_) => Err(RevlogError::AmbiguousPrefix), |
|
277 | 283 | Err(RevlogError::InvalidRevision) => Ok(NULL_REVISION), |
|
278 | 284 | res => res, |
|
279 | 285 | }; |
|
280 | 286 | }; |
|
281 | 287 | |
|
282 | 288 | looked_up |
|
283 | 289 | } |
|
284 | 290 | |
|
285 | 291 | /// Same as `rev_from_node`, without using a persistent nodemap |
|
286 | 292 | /// |
|
287 | 293 | /// This is used as fallback when a persistent nodemap is not present. |
|
288 | 294 | /// This happens when the persistent-nodemap experimental feature is not |
|
289 | 295 | /// enabled, or for small revlogs. |
|
290 | 296 | fn rev_from_node_no_persistent_nodemap( |
|
291 | 297 | &self, |
|
292 | 298 | node: NodePrefix, |
|
293 | 299 | ) -> Result<Revision, RevlogError> { |
|
294 | 300 | // Linear scan of the revlog |
|
295 | 301 | // TODO: consider building a non-persistent nodemap in memory to |
|
296 | 302 | // optimize these cases. |
|
297 | 303 | let mut found_by_prefix = None; |
|
298 | 304 | for rev in (0..self.len() as Revision).rev() { |
|
299 | 305 | let index_entry = self.index.get_entry(rev).ok_or_else(|| { |
|
300 | 306 | HgError::corrupted( |
|
301 | 307 | "revlog references a revision not in the index", |
|
302 | 308 | ) |
|
303 | 309 | })?; |
|
304 | 310 | if node == *index_entry.hash() { |
|
305 | 311 | return Ok(rev); |
|
306 | 312 | } |
|
307 | 313 | if node.is_prefix_of(index_entry.hash()) { |
|
308 | 314 | if found_by_prefix.is_some() { |
|
309 | 315 | return Err(RevlogError::AmbiguousPrefix); |
|
310 | 316 | } |
|
311 | 317 | found_by_prefix = Some(rev) |
|
312 | 318 | } |
|
313 | 319 | } |
|
314 | 320 | found_by_prefix.ok_or(RevlogError::InvalidRevision) |
|
315 | 321 | } |
|
316 | 322 | |
|
317 | 323 | /// Returns whether the given revision exists in this revlog. |
|
318 | 324 | pub fn has_rev(&self, rev: UncheckedRevision) -> bool { |
|
319 | 325 | self.index.check_revision(rev).is_some() |
|
320 | 326 | } |
|
321 | 327 | |
|
322 | 328 | /// Return the full data associated to a revision. |
|
323 | 329 | /// |
|
324 | 330 | /// All entries required to build the final data out of deltas will be |
|
325 | 331 | /// retrieved as needed, and the deltas will be applied to the inital |
|
326 | 332 | /// snapshot to rebuild the final data. |
|
327 | 333 | pub fn get_rev_data( |
|
328 | 334 | &self, |
|
329 | 335 | rev: UncheckedRevision, |
|
330 | 336 | ) -> Result<Cow<[u8]>, RevlogError> { |
|
331 | 337 | if rev == NULL_REVISION.into() { |
|
332 | 338 | return Ok(Cow::Borrowed(&[])); |
|
333 | 339 | }; |
|
334 | 340 | self.get_entry(rev)?.data() |
|
335 | 341 | } |
|
336 | 342 | |
|
337 | 343 | /// [`Self::get_rev_data`] for checked revisions. |
|
338 | 344 | pub fn get_rev_data_for_checked_rev( |
|
339 | 345 | &self, |
|
340 | 346 | rev: Revision, |
|
341 | 347 | ) -> Result<Cow<[u8]>, RevlogError> { |
|
342 | 348 | if rev == NULL_REVISION { |
|
343 | 349 | return Ok(Cow::Borrowed(&[])); |
|
344 | 350 | }; |
|
345 | 351 | self.get_entry_for_checked_rev(rev)?.data() |
|
346 | 352 | } |
|
347 | 353 | |
|
348 | 354 | /// Check the hash of some given data against the recorded hash. |
|
349 | 355 | pub fn check_hash( |
|
350 | 356 | &self, |
|
351 | 357 | p1: Revision, |
|
352 | 358 | p2: Revision, |
|
353 | 359 | expected: &[u8], |
|
354 | 360 | data: &[u8], |
|
355 | 361 | ) -> bool { |
|
356 | 362 | let e1 = self.index.get_entry(p1); |
|
357 | 363 | let h1 = match e1 { |
|
358 | 364 | Some(ref entry) => entry.hash(), |
|
359 | 365 | None => &NULL_NODE, |
|
360 | 366 | }; |
|
361 | 367 | let e2 = self.index.get_entry(p2); |
|
362 | 368 | let h2 = match e2 { |
|
363 | 369 | Some(ref entry) => entry.hash(), |
|
364 | 370 | None => &NULL_NODE, |
|
365 | 371 | }; |
|
366 | 372 | |
|
367 | 373 | hash(data, h1.as_bytes(), h2.as_bytes()) == expected |
|
368 | 374 | } |
|
369 | 375 | |
|
370 | 376 | /// Build the full data of a revision out its snapshot |
|
371 | 377 | /// and its deltas. |
|
372 | 378 | fn build_data_from_deltas( |
|
373 | 379 | snapshot: RevlogEntry, |
|
374 | 380 | deltas: &[RevlogEntry], |
|
375 | 381 | ) -> Result<Vec<u8>, HgError> { |
|
376 | 382 | let snapshot = snapshot.data_chunk()?; |
|
377 | 383 | let deltas = deltas |
|
378 | 384 | .iter() |
|
379 | 385 | .rev() |
|
380 | 386 | .map(RevlogEntry::data_chunk) |
|
381 | 387 | .collect::<Result<Vec<_>, _>>()?; |
|
382 | 388 | let patches: Vec<_> = |
|
383 | 389 | deltas.iter().map(|d| patch::PatchList::new(d)).collect(); |
|
384 | 390 | let patch = patch::fold_patch_lists(&patches); |
|
385 | 391 | Ok(patch.apply(&snapshot)) |
|
386 | 392 | } |
|
387 | 393 | |
|
388 | 394 | /// Return the revlog data. |
|
389 | 395 | fn data(&self) -> &[u8] { |
|
390 | 396 | match &self.data_bytes { |
|
391 | 397 | Some(data_bytes) => data_bytes, |
|
392 | 398 | None => panic!( |
|
393 | 399 | "forgot to load the data or trying to access inline data" |
|
394 | 400 | ), |
|
395 | 401 | } |
|
396 | 402 | } |
|
397 | 403 | |
|
398 | 404 | pub fn make_null_entry(&self) -> RevlogEntry { |
|
399 | 405 | RevlogEntry { |
|
400 | 406 | revlog: self, |
|
401 | 407 | rev: NULL_REVISION, |
|
402 | 408 | bytes: b"", |
|
403 | 409 | compressed_len: 0, |
|
404 | 410 | uncompressed_len: 0, |
|
405 | 411 | base_rev_or_base_of_delta_chain: None, |
|
406 | 412 | p1: NULL_REVISION, |
|
407 | 413 | p2: NULL_REVISION, |
|
408 | 414 | flags: NULL_REVLOG_ENTRY_FLAGS, |
|
409 | 415 | hash: NULL_NODE, |
|
410 | 416 | } |
|
411 | 417 | } |
|
412 | 418 | |
|
413 | 419 | fn get_entry_for_checked_rev( |
|
414 | 420 | &self, |
|
415 | 421 | rev: Revision, |
|
416 | 422 | ) -> Result<RevlogEntry, RevlogError> { |
|
417 | 423 | if rev == NULL_REVISION { |
|
418 | 424 | return Ok(self.make_null_entry()); |
|
419 | 425 | } |
|
420 | 426 | let index_entry = self |
|
421 | 427 | .index |
|
422 | 428 | .get_entry(rev) |
|
423 | 429 | .ok_or(RevlogError::InvalidRevision)?; |
|
424 | 430 | let start = index_entry.offset(); |
|
425 | 431 | let end = start + index_entry.compressed_len() as usize; |
|
426 | 432 | let data = if self.index.is_inline() { |
|
427 | 433 | self.index.data(start, end) |
|
428 | 434 | } else { |
|
429 | 435 | &self.data()[start..end] |
|
430 | 436 | }; |
|
431 | 437 | let base_rev = self |
|
432 | 438 | .index |
|
433 | 439 | .check_revision(index_entry.base_revision_or_base_of_delta_chain()) |
|
434 | 440 | .ok_or_else(|| { |
|
435 | 441 | RevlogError::corrupted(format!( |
|
436 | 442 | "base revision for rev {} is invalid", |
|
437 | 443 | rev |
|
438 | 444 | )) |
|
439 | 445 | })?; |
|
440 | 446 | let p1 = |
|
441 | 447 | self.index.check_revision(index_entry.p1()).ok_or_else(|| { |
|
442 | 448 | RevlogError::corrupted(format!( |
|
443 | 449 | "p1 for rev {} is invalid", |
|
444 | 450 | rev |
|
445 | 451 | )) |
|
446 | 452 | })?; |
|
447 | 453 | let p2 = |
|
448 | 454 | self.index.check_revision(index_entry.p2()).ok_or_else(|| { |
|
449 | 455 | RevlogError::corrupted(format!( |
|
450 | 456 | "p2 for rev {} is invalid", |
|
451 | 457 | rev |
|
452 | 458 | )) |
|
453 | 459 | })?; |
|
454 | 460 | let entry = RevlogEntry { |
|
455 | 461 | revlog: self, |
|
456 | 462 | rev, |
|
457 | 463 | bytes: data, |
|
458 | 464 | compressed_len: index_entry.compressed_len(), |
|
459 | 465 | uncompressed_len: index_entry.uncompressed_len(), |
|
460 | 466 | base_rev_or_base_of_delta_chain: if base_rev == rev { |
|
461 | 467 | None |
|
462 | 468 | } else { |
|
463 | 469 | Some(base_rev) |
|
464 | 470 | }, |
|
465 | 471 | p1, |
|
466 | 472 | p2, |
|
467 | 473 | flags: index_entry.flags(), |
|
468 | 474 | hash: *index_entry.hash(), |
|
469 | 475 | }; |
|
470 | 476 | Ok(entry) |
|
471 | 477 | } |
|
472 | 478 | |
|
473 | 479 | /// Get an entry of the revlog. |
|
474 | 480 | pub fn get_entry( |
|
475 | 481 | &self, |
|
476 | 482 | rev: UncheckedRevision, |
|
477 | 483 | ) -> Result<RevlogEntry, RevlogError> { |
|
478 | 484 | if rev == NULL_REVISION.into() { |
|
479 | 485 | return Ok(self.make_null_entry()); |
|
480 | 486 | } |
|
481 | 487 | let rev = self.index.check_revision(rev).ok_or_else(|| { |
|
482 | 488 | RevlogError::corrupted(format!("rev {} is invalid", rev)) |
|
483 | 489 | })?; |
|
484 | 490 | self.get_entry_for_checked_rev(rev) |
|
485 | 491 | } |
|
486 | 492 | } |
|
487 | 493 | |
|
488 | 494 | /// The revlog entry's bytes and the necessary informations to extract |
|
489 | 495 | /// the entry's data. |
|
490 | 496 | #[derive(Clone)] |
|
491 | 497 | pub struct RevlogEntry<'revlog> { |
|
492 | 498 | revlog: &'revlog Revlog, |
|
493 | 499 | rev: Revision, |
|
494 | 500 | bytes: &'revlog [u8], |
|
495 | 501 | compressed_len: u32, |
|
496 | 502 | uncompressed_len: i32, |
|
497 | 503 | base_rev_or_base_of_delta_chain: Option<Revision>, |
|
498 | 504 | p1: Revision, |
|
499 | 505 | p2: Revision, |
|
500 | 506 | flags: u16, |
|
501 | 507 | hash: Node, |
|
502 | 508 | } |
|
503 | 509 | |
|
504 | 510 | thread_local! { |
|
505 | 511 | // seems fine to [unwrap] here: this can only fail due to memory allocation |
|
506 | 512 | // failing, and it's normal for that to cause panic. |
|
507 | 513 | static ZSTD_DECODER : RefCell<zstd::bulk::Decompressor<'static>> = |
|
508 | 514 | RefCell::new(zstd::bulk::Decompressor::new().ok().unwrap()); |
|
509 | 515 | } |
|
510 | 516 | |
|
511 | 517 | fn zstd_decompress_to_buffer( |
|
512 | 518 | bytes: &[u8], |
|
513 | 519 | buf: &mut Vec<u8>, |
|
514 | 520 | ) -> Result<usize, std::io::Error> { |
|
515 | 521 | ZSTD_DECODER |
|
516 | 522 | .with(|decoder| decoder.borrow_mut().decompress_to_buffer(bytes, buf)) |
|
517 | 523 | } |
|
518 | 524 | |
|
519 | 525 | impl<'revlog> RevlogEntry<'revlog> { |
|
520 | 526 | pub fn revision(&self) -> Revision { |
|
521 | 527 | self.rev |
|
522 | 528 | } |
|
523 | 529 | |
|
524 | 530 | pub fn node(&self) -> &Node { |
|
525 | 531 | &self.hash |
|
526 | 532 | } |
|
527 | 533 | |
|
528 | 534 | pub fn uncompressed_len(&self) -> Option<u32> { |
|
529 | 535 | u32::try_from(self.uncompressed_len).ok() |
|
530 | 536 | } |
|
531 | 537 | |
|
532 | 538 | pub fn has_p1(&self) -> bool { |
|
533 | 539 | self.p1 != NULL_REVISION |
|
534 | 540 | } |
|
535 | 541 | |
|
536 | 542 | pub fn p1_entry( |
|
537 | 543 | &self, |
|
538 | 544 | ) -> Result<Option<RevlogEntry<'revlog>>, RevlogError> { |
|
539 | 545 | if self.p1 == NULL_REVISION { |
|
540 | 546 | Ok(None) |
|
541 | 547 | } else { |
|
542 | 548 | Ok(Some(self.revlog.get_entry_for_checked_rev(self.p1)?)) |
|
543 | 549 | } |
|
544 | 550 | } |
|
545 | 551 | |
|
546 | 552 | pub fn p2_entry( |
|
547 | 553 | &self, |
|
548 | 554 | ) -> Result<Option<RevlogEntry<'revlog>>, RevlogError> { |
|
549 | 555 | if self.p2 == NULL_REVISION { |
|
550 | 556 | Ok(None) |
|
551 | 557 | } else { |
|
552 | 558 | Ok(Some(self.revlog.get_entry_for_checked_rev(self.p2)?)) |
|
553 | 559 | } |
|
554 | 560 | } |
|
555 | 561 | |
|
556 | 562 | pub fn p1(&self) -> Option<Revision> { |
|
557 | 563 | if self.p1 == NULL_REVISION { |
|
558 | 564 | None |
|
559 | 565 | } else { |
|
560 | 566 | Some(self.p1) |
|
561 | 567 | } |
|
562 | 568 | } |
|
563 | 569 | |
|
564 | 570 | pub fn p2(&self) -> Option<Revision> { |
|
565 | 571 | if self.p2 == NULL_REVISION { |
|
566 | 572 | None |
|
567 | 573 | } else { |
|
568 | 574 | Some(self.p2) |
|
569 | 575 | } |
|
570 | 576 | } |
|
571 | 577 | |
|
572 | 578 | pub fn is_censored(&self) -> bool { |
|
573 | 579 | (self.flags & REVISION_FLAG_CENSORED) != 0 |
|
574 | 580 | } |
|
575 | 581 | |
|
576 | 582 | pub fn has_length_affecting_flag_processor(&self) -> bool { |
|
577 | 583 | // Relevant Python code: revlog.size() |
|
578 | 584 | // note: ELLIPSIS is known to not change the content |
|
579 | 585 | (self.flags & (REVIDX_KNOWN_FLAGS ^ REVISION_FLAG_ELLIPSIS)) != 0 |
|
580 | 586 | } |
|
581 | 587 | |
|
582 | 588 | /// The data for this entry, after resolving deltas if any. |
|
583 | 589 | pub fn rawdata(&self) -> Result<Cow<'revlog, [u8]>, RevlogError> { |
|
584 | 590 | let mut entry = self.clone(); |
|
585 | 591 | let mut delta_chain = vec![]; |
|
586 | 592 | |
|
587 | 593 | // The meaning of `base_rev_or_base_of_delta_chain` depends on |
|
588 | 594 | // generaldelta. See the doc on `ENTRY_DELTA_BASE` in |
|
589 | 595 | // `mercurial/revlogutils/constants.py` and the code in |
|
590 | 596 | // [_chaininfo] and in [index_deltachain]. |
|
591 | 597 | let uses_generaldelta = self.revlog.index.uses_generaldelta(); |
|
592 | 598 | while let Some(base_rev) = entry.base_rev_or_base_of_delta_chain { |
|
593 | 599 | entry = if uses_generaldelta { |
|
594 | 600 | delta_chain.push(entry); |
|
595 | 601 | self.revlog.get_entry_for_checked_rev(base_rev)? |
|
596 | 602 | } else { |
|
597 | 603 | let base_rev = UncheckedRevision(entry.rev - 1); |
|
598 | 604 | delta_chain.push(entry); |
|
599 | 605 | self.revlog.get_entry(base_rev)? |
|
600 | 606 | }; |
|
601 | 607 | } |
|
602 | 608 | |
|
603 | 609 | let data = if delta_chain.is_empty() { |
|
604 | 610 | entry.data_chunk()? |
|
605 | 611 | } else { |
|
606 | 612 | Revlog::build_data_from_deltas(entry, &delta_chain)?.into() |
|
607 | 613 | }; |
|
608 | 614 | |
|
609 | 615 | Ok(data) |
|
610 | 616 | } |
|
611 | 617 | |
|
612 | 618 | fn check_data( |
|
613 | 619 | &self, |
|
614 | 620 | data: Cow<'revlog, [u8]>, |
|
615 | 621 | ) -> Result<Cow<'revlog, [u8]>, RevlogError> { |
|
616 | 622 | if self.revlog.check_hash( |
|
617 | 623 | self.p1, |
|
618 | 624 | self.p2, |
|
619 | 625 | self.hash.as_bytes(), |
|
620 | 626 | &data, |
|
621 | 627 | ) { |
|
622 | 628 | Ok(data) |
|
623 | 629 | } else { |
|
624 | 630 | if (self.flags & REVISION_FLAG_ELLIPSIS) != 0 { |
|
625 | 631 | return Err(HgError::unsupported( |
|
626 | 632 | "ellipsis revisions are not supported by rhg", |
|
627 | 633 | ) |
|
628 | 634 | .into()); |
|
629 | 635 | } |
|
630 | 636 | Err(corrupted(format!( |
|
631 | 637 | "hash check failed for revision {}", |
|
632 | 638 | self.rev |
|
633 | 639 | )) |
|
634 | 640 | .into()) |
|
635 | 641 | } |
|
636 | 642 | } |
|
637 | 643 | |
|
638 | 644 | pub fn data(&self) -> Result<Cow<'revlog, [u8]>, RevlogError> { |
|
639 | 645 | let data = self.rawdata()?; |
|
640 | 646 | if self.rev == NULL_REVISION { |
|
641 | 647 | return Ok(data); |
|
642 | 648 | } |
|
643 | 649 | if self.is_censored() { |
|
644 | 650 | return Err(HgError::CensoredNodeError.into()); |
|
645 | 651 | } |
|
646 | 652 | self.check_data(data) |
|
647 | 653 | } |
|
648 | 654 | |
|
649 | 655 | /// Extract the data contained in the entry. |
|
650 | 656 | /// This may be a delta. (See `is_delta`.) |
|
651 | 657 | fn data_chunk(&self) -> Result<Cow<'revlog, [u8]>, HgError> { |
|
652 | 658 | if self.bytes.is_empty() { |
|
653 | 659 | return Ok(Cow::Borrowed(&[])); |
|
654 | 660 | } |
|
655 | 661 | match self.bytes[0] { |
|
656 | 662 | // Revision data is the entirety of the entry, including this |
|
657 | 663 | // header. |
|
658 | 664 | b'\0' => Ok(Cow::Borrowed(self.bytes)), |
|
659 | 665 | // Raw revision data follows. |
|
660 | 666 | b'u' => Ok(Cow::Borrowed(&self.bytes[1..])), |
|
661 | 667 | // zlib (RFC 1950) data. |
|
662 | 668 | b'x' => Ok(Cow::Owned(self.uncompressed_zlib_data()?)), |
|
663 | 669 | // zstd data. |
|
664 | 670 | b'\x28' => Ok(Cow::Owned(self.uncompressed_zstd_data()?)), |
|
665 | 671 | // A proper new format should have had a repo/store requirement. |
|
666 | 672 | format_type => Err(corrupted(format!( |
|
667 | 673 | "unknown compression header '{}'", |
|
668 | 674 | format_type |
|
669 | 675 | ))), |
|
670 | 676 | } |
|
671 | 677 | } |
|
672 | 678 | |
|
673 | 679 | fn uncompressed_zlib_data(&self) -> Result<Vec<u8>, HgError> { |
|
674 | 680 | let mut decoder = ZlibDecoder::new(self.bytes); |
|
675 | 681 | if self.is_delta() { |
|
676 | 682 | let mut buf = Vec::with_capacity(self.compressed_len as usize); |
|
677 | 683 | decoder |
|
678 | 684 | .read_to_end(&mut buf) |
|
679 | 685 | .map_err(|e| corrupted(e.to_string()))?; |
|
680 | 686 | Ok(buf) |
|
681 | 687 | } else { |
|
682 | 688 | let cap = self.uncompressed_len.max(0) as usize; |
|
683 | 689 | let mut buf = vec![0; cap]; |
|
684 | 690 | decoder |
|
685 | 691 | .read_exact(&mut buf) |
|
686 | 692 | .map_err(|e| corrupted(e.to_string()))?; |
|
687 | 693 | Ok(buf) |
|
688 | 694 | } |
|
689 | 695 | } |
|
690 | 696 | |
|
691 | 697 | fn uncompressed_zstd_data(&self) -> Result<Vec<u8>, HgError> { |
|
692 | 698 | let cap = self.uncompressed_len.max(0) as usize; |
|
693 | 699 | if self.is_delta() { |
|
694 | 700 | // [cap] is usually an over-estimate of the space needed because |
|
695 | 701 | // it's the length of delta-decoded data, but we're interested |
|
696 | 702 | // in the size of the delta. |
|
697 | 703 | // This means we have to [shrink_to_fit] to avoid holding on |
|
698 | 704 | // to a large chunk of memory, but it also means we must have a |
|
699 | 705 | // fallback branch, for the case when the delta is longer than |
|
700 | 706 | // the original data (surprisingly, this does happen in practice) |
|
701 | 707 | let mut buf = Vec::with_capacity(cap); |
|
702 | 708 | match zstd_decompress_to_buffer(self.bytes, &mut buf) { |
|
703 | 709 | Ok(_) => buf.shrink_to_fit(), |
|
704 | 710 | Err(_) => { |
|
705 | 711 | buf.clear(); |
|
706 | 712 | zstd::stream::copy_decode(self.bytes, &mut buf) |
|
707 | 713 | .map_err(|e| corrupted(e.to_string()))?; |
|
708 | 714 | } |
|
709 | 715 | }; |
|
710 | 716 | Ok(buf) |
|
711 | 717 | } else { |
|
712 | 718 | let mut buf = Vec::with_capacity(cap); |
|
713 | 719 | let len = zstd_decompress_to_buffer(self.bytes, &mut buf) |
|
714 | 720 | .map_err(|e| corrupted(e.to_string()))?; |
|
715 | 721 | if len != self.uncompressed_len as usize { |
|
716 | 722 | Err(corrupted("uncompressed length does not match")) |
|
717 | 723 | } else { |
|
718 | 724 | Ok(buf) |
|
719 | 725 | } |
|
720 | 726 | } |
|
721 | 727 | } |
|
722 | 728 | |
|
723 | 729 | /// Tell if the entry is a snapshot or a delta |
|
724 | 730 | /// (influences on decompression). |
|
725 | 731 | fn is_delta(&self) -> bool { |
|
726 | 732 | self.base_rev_or_base_of_delta_chain.is_some() |
|
727 | 733 | } |
|
728 | 734 | } |
|
729 | 735 | |
|
730 | 736 | /// Calculate the hash of a revision given its data and its parents. |
|
731 | 737 | fn hash( |
|
732 | 738 | data: &[u8], |
|
733 | 739 | p1_hash: &[u8], |
|
734 | 740 | p2_hash: &[u8], |
|
735 | 741 | ) -> [u8; NODE_BYTES_LENGTH] { |
|
736 | 742 | let mut hasher = Sha1::new(); |
|
737 | 743 | let (a, b) = (p1_hash, p2_hash); |
|
738 | 744 | if a > b { |
|
739 | 745 | hasher.update(b); |
|
740 | 746 | hasher.update(a); |
|
741 | 747 | } else { |
|
742 | 748 | hasher.update(a); |
|
743 | 749 | hasher.update(b); |
|
744 | 750 | } |
|
745 | 751 | hasher.update(data); |
|
746 | 752 | *hasher.finalize().as_ref() |
|
747 | 753 | } |
|
748 | 754 | |
|
749 | 755 | #[cfg(test)] |
|
750 | 756 | mod tests { |
|
751 | 757 | use super::*; |
|
752 | 758 | use crate::index::{IndexEntryBuilder, INDEX_ENTRY_SIZE}; |
|
753 | 759 | use itertools::Itertools; |
|
754 | 760 | |
|
755 | 761 | #[test] |
|
756 | 762 | fn test_empty() { |
|
757 | 763 | let temp = tempfile::tempdir().unwrap(); |
|
758 | 764 | let vfs = Vfs { base: temp.path() }; |
|
759 | 765 | std::fs::write(temp.path().join("foo.i"), b"").unwrap(); |
|
760 | 766 | let revlog = Revlog::open(&vfs, "foo.i", None, false).unwrap(); |
|
761 | 767 | assert!(revlog.is_empty()); |
|
762 | 768 | assert_eq!(revlog.len(), 0); |
|
763 | 769 | assert!(revlog.get_entry(0.into()).is_err()); |
|
764 | 770 | assert!(!revlog.has_rev(0.into())); |
|
765 | 771 | assert_eq!( |
|
766 | 772 | revlog.rev_from_node(NULL_NODE.into()).unwrap(), |
|
767 | 773 | NULL_REVISION |
|
768 | 774 | ); |
|
769 | 775 | let null_entry = revlog.get_entry(NULL_REVISION.into()).ok().unwrap(); |
|
770 | 776 | assert_eq!(null_entry.revision(), NULL_REVISION); |
|
771 | 777 | assert!(null_entry.data().unwrap().is_empty()); |
|
772 | 778 | } |
|
773 | 779 | |
|
774 | 780 | #[test] |
|
775 | 781 | fn test_inline() { |
|
776 | 782 | let temp = tempfile::tempdir().unwrap(); |
|
777 | 783 | let vfs = Vfs { base: temp.path() }; |
|
778 | 784 | let node0 = Node::from_hex("2ed2a3912a0b24502043eae84ee4b279c18b90dd") |
|
779 | 785 | .unwrap(); |
|
780 | 786 | let node1 = Node::from_hex("b004912a8510032a0350a74daa2803dadfb00e12") |
|
781 | 787 | .unwrap(); |
|
782 | 788 | let node2 = Node::from_hex("dd6ad206e907be60927b5a3117b97dffb2590582") |
|
783 | 789 | .unwrap(); |
|
784 | 790 | let entry0_bytes = IndexEntryBuilder::new() |
|
785 | 791 | .is_first(true) |
|
786 | 792 | .with_version(1) |
|
787 | 793 | .with_inline(true) |
|
788 | 794 | .with_offset(INDEX_ENTRY_SIZE) |
|
789 | 795 | .with_node(node0) |
|
790 | 796 | .build(); |
|
791 | 797 | let entry1_bytes = IndexEntryBuilder::new() |
|
792 | 798 | .with_offset(INDEX_ENTRY_SIZE) |
|
793 | 799 | .with_node(node1) |
|
794 | 800 | .build(); |
|
795 | 801 | let entry2_bytes = IndexEntryBuilder::new() |
|
796 | 802 | .with_offset(INDEX_ENTRY_SIZE) |
|
797 | 803 | .with_p1(0) |
|
798 | 804 | .with_p2(1) |
|
799 | 805 | .with_node(node2) |
|
800 | 806 | .build(); |
|
801 | 807 | let contents = vec![entry0_bytes, entry1_bytes, entry2_bytes] |
|
802 | 808 | .into_iter() |
|
803 | 809 | .flatten() |
|
804 | 810 | .collect_vec(); |
|
805 | 811 | std::fs::write(temp.path().join("foo.i"), contents).unwrap(); |
|
806 | 812 | let revlog = Revlog::open(&vfs, "foo.i", None, false).unwrap(); |
|
807 | 813 | |
|
808 | 814 | let entry0 = revlog.get_entry(0.into()).ok().unwrap(); |
|
809 | 815 | assert_eq!(entry0.revision(), 0); |
|
810 | 816 | assert_eq!(*entry0.node(), node0); |
|
811 | 817 | assert!(!entry0.has_p1()); |
|
812 | 818 | assert_eq!(entry0.p1(), None); |
|
813 | 819 | assert_eq!(entry0.p2(), None); |
|
814 | 820 | let p1_entry = entry0.p1_entry().unwrap(); |
|
815 | 821 | assert!(p1_entry.is_none()); |
|
816 | 822 | let p2_entry = entry0.p2_entry().unwrap(); |
|
817 | 823 | assert!(p2_entry.is_none()); |
|
818 | 824 | |
|
819 | 825 | let entry1 = revlog.get_entry(1.into()).ok().unwrap(); |
|
820 | 826 | assert_eq!(entry1.revision(), 1); |
|
821 | 827 | assert_eq!(*entry1.node(), node1); |
|
822 | 828 | assert!(!entry1.has_p1()); |
|
823 | 829 | assert_eq!(entry1.p1(), None); |
|
824 | 830 | assert_eq!(entry1.p2(), None); |
|
825 | 831 | let p1_entry = entry1.p1_entry().unwrap(); |
|
826 | 832 | assert!(p1_entry.is_none()); |
|
827 | 833 | let p2_entry = entry1.p2_entry().unwrap(); |
|
828 | 834 | assert!(p2_entry.is_none()); |
|
829 | 835 | |
|
830 | 836 | let entry2 = revlog.get_entry(2.into()).ok().unwrap(); |
|
831 | 837 | assert_eq!(entry2.revision(), 2); |
|
832 | 838 | assert_eq!(*entry2.node(), node2); |
|
833 | 839 | assert!(entry2.has_p1()); |
|
834 | 840 | assert_eq!(entry2.p1(), Some(0)); |
|
835 | 841 | assert_eq!(entry2.p2(), Some(1)); |
|
836 | 842 | let p1_entry = entry2.p1_entry().unwrap(); |
|
837 | 843 | assert!(p1_entry.is_some()); |
|
838 | 844 | assert_eq!(p1_entry.unwrap().revision(), 0); |
|
839 | 845 | let p2_entry = entry2.p2_entry().unwrap(); |
|
840 | 846 | assert!(p2_entry.is_some()); |
|
841 | 847 | assert_eq!(p2_entry.unwrap().revision(), 1); |
|
842 | 848 | } |
|
843 | 849 | |
|
844 | 850 | #[test] |
|
845 | 851 | fn test_nodemap() { |
|
846 | 852 | let temp = tempfile::tempdir().unwrap(); |
|
847 | 853 | let vfs = Vfs { base: temp.path() }; |
|
848 | 854 | |
|
849 | 855 | // building a revlog with a forced Node starting with zeros |
|
850 | 856 | // This is a corruption, but it does not preclude using the nodemap |
|
851 | 857 | // if we don't try and access the data |
|
852 | 858 | let node0 = Node::from_hex("00d2a3912a0b24502043eae84ee4b279c18b90dd") |
|
853 | 859 | .unwrap(); |
|
854 | 860 | let node1 = Node::from_hex("b004912a8510032a0350a74daa2803dadfb00e12") |
|
855 | 861 | .unwrap(); |
|
856 | 862 | let entry0_bytes = IndexEntryBuilder::new() |
|
857 | 863 | .is_first(true) |
|
858 | 864 | .with_version(1) |
|
859 | 865 | .with_inline(true) |
|
860 | 866 | .with_offset(INDEX_ENTRY_SIZE) |
|
861 | 867 | .with_node(node0) |
|
862 | 868 | .build(); |
|
863 | 869 | let entry1_bytes = IndexEntryBuilder::new() |
|
864 | 870 | .with_offset(INDEX_ENTRY_SIZE) |
|
865 | 871 | .with_node(node1) |
|
866 | 872 | .build(); |
|
867 | 873 | let contents = vec![entry0_bytes, entry1_bytes] |
|
868 | 874 | .into_iter() |
|
869 | 875 | .flatten() |
|
870 | 876 | .collect_vec(); |
|
871 | 877 | std::fs::write(temp.path().join("foo.i"), contents).unwrap(); |
|
872 | 878 | let revlog = Revlog::open(&vfs, "foo.i", None, false).unwrap(); |
|
873 | 879 | |
|
874 | 880 | // accessing the data shows the corruption |
|
875 | 881 | revlog.get_entry(0.into()).unwrap().data().unwrap_err(); |
|
876 | 882 | |
|
877 | 883 | assert_eq!(revlog.rev_from_node(NULL_NODE.into()).unwrap(), -1); |
|
878 | 884 | assert_eq!(revlog.rev_from_node(node0.into()).unwrap(), 0); |
|
879 | 885 | assert_eq!(revlog.rev_from_node(node1.into()).unwrap(), 1); |
|
880 | 886 | assert_eq!( |
|
881 | 887 | revlog |
|
882 | 888 | .rev_from_node(NodePrefix::from_hex("000").unwrap()) |
|
883 | 889 | .unwrap(), |
|
884 | 890 | -1 |
|
885 | 891 | ); |
|
886 | 892 | assert_eq!( |
|
887 | 893 | revlog |
|
888 | 894 | .rev_from_node(NodePrefix::from_hex("b00").unwrap()) |
|
889 | 895 | .unwrap(), |
|
890 | 896 | 1 |
|
891 | 897 | ); |
|
892 | 898 | // RevlogError does not implement PartialEq |
|
893 | 899 | // (ultimately because io::Error does not) |
|
894 | 900 | match revlog |
|
895 | 901 | .rev_from_node(NodePrefix::from_hex("00").unwrap()) |
|
896 | 902 | .expect_err("Expected to give AmbiguousPrefix error") |
|
897 | 903 | { |
|
898 | 904 | RevlogError::AmbiguousPrefix => (), |
|
899 | 905 | e => { |
|
900 | 906 | panic!("Got another error than AmbiguousPrefix: {:?}", e); |
|
901 | 907 | } |
|
902 | 908 | }; |
|
903 | 909 | } |
|
904 | 910 | } |
General Comments 0
You need to be logged in to leave comments.
Login now