Show More
@@ -1,1203 +1,1213 b'' | |||||
1 | use bytes_cast::BytesCast; |
|
1 | use bytes_cast::BytesCast; | |
2 | use micro_timer::timed; |
|
2 | use micro_timer::timed; | |
3 | use std::borrow::Cow; |
|
3 | use std::borrow::Cow; | |
4 | use std::path::PathBuf; |
|
4 | use std::path::PathBuf; | |
5 |
|
5 | |||
6 | use super::on_disk; |
|
6 | use super::on_disk; | |
7 | use super::on_disk::DirstateV2ParseError; |
|
7 | use super::on_disk::DirstateV2ParseError; | |
8 | use super::owning::OwningDirstateMap; |
|
8 | use super::owning::OwningDirstateMap; | |
9 | use super::path_with_basename::WithBasename; |
|
9 | use super::path_with_basename::WithBasename; | |
10 | use crate::dirstate::parsers::pack_entry; |
|
10 | use crate::dirstate::parsers::pack_entry; | |
11 | use crate::dirstate::parsers::packed_entry_size; |
|
11 | use crate::dirstate::parsers::packed_entry_size; | |
12 | use crate::dirstate::parsers::parse_dirstate_entries; |
|
12 | use crate::dirstate::parsers::parse_dirstate_entries; | |
13 | use crate::dirstate::CopyMapIter; |
|
13 | use crate::dirstate::CopyMapIter; | |
14 | use crate::dirstate::StateMapIter; |
|
14 | use crate::dirstate::StateMapIter; | |
15 | use crate::dirstate::TruncatedTimestamp; |
|
15 | use crate::dirstate::TruncatedTimestamp; | |
16 | use crate::dirstate::SIZE_FROM_OTHER_PARENT; |
|
16 | use crate::dirstate::SIZE_FROM_OTHER_PARENT; | |
17 | use crate::dirstate::SIZE_NON_NORMAL; |
|
17 | use crate::dirstate::SIZE_NON_NORMAL; | |
18 | use crate::matchers::Matcher; |
|
18 | use crate::matchers::Matcher; | |
19 | use crate::utils::hg_path::{HgPath, HgPathBuf}; |
|
19 | use crate::utils::hg_path::{HgPath, HgPathBuf}; | |
20 | use crate::DirstateEntry; |
|
20 | use crate::DirstateEntry; | |
21 | use crate::DirstateError; |
|
21 | use crate::DirstateError; | |
22 | use crate::DirstateParents; |
|
22 | use crate::DirstateParents; | |
23 | use crate::DirstateStatus; |
|
23 | use crate::DirstateStatus; | |
24 | use crate::EntryState; |
|
24 | use crate::EntryState; | |
25 | use crate::FastHashMap; |
|
25 | use crate::FastHashMap; | |
26 | use crate::PatternFileWarning; |
|
26 | use crate::PatternFileWarning; | |
27 | use crate::StatusError; |
|
27 | use crate::StatusError; | |
28 | use crate::StatusOptions; |
|
28 | use crate::StatusOptions; | |
29 |
|
29 | |||
30 | /// Append to an existing data file if the amount of unreachable data (not used |
|
30 | /// Append to an existing data file if the amount of unreachable data (not used | |
31 | /// anymore) is less than this fraction of the total amount of existing data. |
|
31 | /// anymore) is less than this fraction of the total amount of existing data. | |
32 | const ACCEPTABLE_UNREACHABLE_BYTES_RATIO: f32 = 0.5; |
|
32 | const ACCEPTABLE_UNREACHABLE_BYTES_RATIO: f32 = 0.5; | |
33 |
|
33 | |||
|
34 | #[derive(Debug, PartialEq, Eq)] | |||
|
35 | /// Version of the on-disk format | |||
|
36 | pub enum DirstateVersion { | |||
|
37 | V1, | |||
|
38 | V2, | |||
|
39 | } | |||
|
40 | ||||
34 | pub struct DirstateMap<'on_disk> { |
|
41 | pub struct DirstateMap<'on_disk> { | |
35 | /// Contents of the `.hg/dirstate` file |
|
42 | /// Contents of the `.hg/dirstate` file | |
36 | pub(super) on_disk: &'on_disk [u8], |
|
43 | pub(super) on_disk: &'on_disk [u8], | |
37 |
|
44 | |||
38 | pub(super) root: ChildNodes<'on_disk>, |
|
45 | pub(super) root: ChildNodes<'on_disk>, | |
39 |
|
46 | |||
40 | /// Number of nodes anywhere in the tree that have `.entry.is_some()`. |
|
47 | /// Number of nodes anywhere in the tree that have `.entry.is_some()`. | |
41 | pub(super) nodes_with_entry_count: u32, |
|
48 | pub(super) nodes_with_entry_count: u32, | |
42 |
|
49 | |||
43 | /// Number of nodes anywhere in the tree that have |
|
50 | /// Number of nodes anywhere in the tree that have | |
44 | /// `.copy_source.is_some()`. |
|
51 | /// `.copy_source.is_some()`. | |
45 | pub(super) nodes_with_copy_source_count: u32, |
|
52 | pub(super) nodes_with_copy_source_count: u32, | |
46 |
|
53 | |||
47 | /// See on_disk::Header |
|
54 | /// See on_disk::Header | |
48 | pub(super) ignore_patterns_hash: on_disk::IgnorePatternsHash, |
|
55 | pub(super) ignore_patterns_hash: on_disk::IgnorePatternsHash, | |
49 |
|
56 | |||
50 | /// How many bytes of `on_disk` are not used anymore |
|
57 | /// How many bytes of `on_disk` are not used anymore | |
51 | pub(super) unreachable_bytes: u32, |
|
58 | pub(super) unreachable_bytes: u32, | |
52 |
|
59 | |||
53 | /// Size of the data used to first load this `DirstateMap`. Used in case |
|
60 | /// Size of the data used to first load this `DirstateMap`. Used in case | |
54 | /// we need to write some new metadata, but no new data on disk. |
|
61 | /// we need to write some new metadata, but no new data on disk. | |
55 | pub(super) old_data_size: usize, |
|
62 | pub(super) old_data_size: usize, | |
|
63 | ||||
|
64 | pub(super) dirstate_version: DirstateVersion, | |||
56 | } |
|
65 | } | |
57 |
|
66 | |||
58 | /// Using a plain `HgPathBuf` of the full path from the repository root as a |
|
67 | /// Using a plain `HgPathBuf` of the full path from the repository root as a | |
59 | /// map key would also work: all paths in a given map have the same parent |
|
68 | /// map key would also work: all paths in a given map have the same parent | |
60 | /// path, so comparing full paths gives the same result as comparing base |
|
69 | /// path, so comparing full paths gives the same result as comparing base | |
61 | /// names. However `HashMap` would waste time always re-hashing the same |
|
70 | /// names. However `HashMap` would waste time always re-hashing the same | |
62 | /// string prefix. |
|
71 | /// string prefix. | |
63 | pub(super) type NodeKey<'on_disk> = WithBasename<Cow<'on_disk, HgPath>>; |
|
72 | pub(super) type NodeKey<'on_disk> = WithBasename<Cow<'on_disk, HgPath>>; | |
64 |
|
73 | |||
65 | /// Similar to `&'tree Cow<'on_disk, HgPath>`, but can also be returned |
|
74 | /// Similar to `&'tree Cow<'on_disk, HgPath>`, but can also be returned | |
66 | /// for on-disk nodes that don’t actually have a `Cow` to borrow. |
|
75 | /// for on-disk nodes that don’t actually have a `Cow` to borrow. | |
67 | pub(super) enum BorrowedPath<'tree, 'on_disk> { |
|
76 | pub(super) enum BorrowedPath<'tree, 'on_disk> { | |
68 | InMemory(&'tree HgPathBuf), |
|
77 | InMemory(&'tree HgPathBuf), | |
69 | OnDisk(&'on_disk HgPath), |
|
78 | OnDisk(&'on_disk HgPath), | |
70 | } |
|
79 | } | |
71 |
|
80 | |||
72 | pub(super) enum ChildNodes<'on_disk> { |
|
81 | pub(super) enum ChildNodes<'on_disk> { | |
73 | InMemory(FastHashMap<NodeKey<'on_disk>, Node<'on_disk>>), |
|
82 | InMemory(FastHashMap<NodeKey<'on_disk>, Node<'on_disk>>), | |
74 | OnDisk(&'on_disk [on_disk::Node]), |
|
83 | OnDisk(&'on_disk [on_disk::Node]), | |
75 | } |
|
84 | } | |
76 |
|
85 | |||
77 | pub(super) enum ChildNodesRef<'tree, 'on_disk> { |
|
86 | pub(super) enum ChildNodesRef<'tree, 'on_disk> { | |
78 | InMemory(&'tree FastHashMap<NodeKey<'on_disk>, Node<'on_disk>>), |
|
87 | InMemory(&'tree FastHashMap<NodeKey<'on_disk>, Node<'on_disk>>), | |
79 | OnDisk(&'on_disk [on_disk::Node]), |
|
88 | OnDisk(&'on_disk [on_disk::Node]), | |
80 | } |
|
89 | } | |
81 |
|
90 | |||
82 | pub(super) enum NodeRef<'tree, 'on_disk> { |
|
91 | pub(super) enum NodeRef<'tree, 'on_disk> { | |
83 | InMemory(&'tree NodeKey<'on_disk>, &'tree Node<'on_disk>), |
|
92 | InMemory(&'tree NodeKey<'on_disk>, &'tree Node<'on_disk>), | |
84 | OnDisk(&'on_disk on_disk::Node), |
|
93 | OnDisk(&'on_disk on_disk::Node), | |
85 | } |
|
94 | } | |
86 |
|
95 | |||
87 | impl<'tree, 'on_disk> BorrowedPath<'tree, 'on_disk> { |
|
96 | impl<'tree, 'on_disk> BorrowedPath<'tree, 'on_disk> { | |
88 | pub fn detach_from_tree(&self) -> Cow<'on_disk, HgPath> { |
|
97 | pub fn detach_from_tree(&self) -> Cow<'on_disk, HgPath> { | |
89 | match *self { |
|
98 | match *self { | |
90 | BorrowedPath::InMemory(in_memory) => Cow::Owned(in_memory.clone()), |
|
99 | BorrowedPath::InMemory(in_memory) => Cow::Owned(in_memory.clone()), | |
91 | BorrowedPath::OnDisk(on_disk) => Cow::Borrowed(on_disk), |
|
100 | BorrowedPath::OnDisk(on_disk) => Cow::Borrowed(on_disk), | |
92 | } |
|
101 | } | |
93 | } |
|
102 | } | |
94 | } |
|
103 | } | |
95 |
|
104 | |||
96 | impl<'tree, 'on_disk> std::ops::Deref for BorrowedPath<'tree, 'on_disk> { |
|
105 | impl<'tree, 'on_disk> std::ops::Deref for BorrowedPath<'tree, 'on_disk> { | |
97 | type Target = HgPath; |
|
106 | type Target = HgPath; | |
98 |
|
107 | |||
99 | fn deref(&self) -> &HgPath { |
|
108 | fn deref(&self) -> &HgPath { | |
100 | match *self { |
|
109 | match *self { | |
101 | BorrowedPath::InMemory(in_memory) => in_memory, |
|
110 | BorrowedPath::InMemory(in_memory) => in_memory, | |
102 | BorrowedPath::OnDisk(on_disk) => on_disk, |
|
111 | BorrowedPath::OnDisk(on_disk) => on_disk, | |
103 | } |
|
112 | } | |
104 | } |
|
113 | } | |
105 | } |
|
114 | } | |
106 |
|
115 | |||
107 | impl Default for ChildNodes<'_> { |
|
116 | impl Default for ChildNodes<'_> { | |
108 | fn default() -> Self { |
|
117 | fn default() -> Self { | |
109 | ChildNodes::InMemory(Default::default()) |
|
118 | ChildNodes::InMemory(Default::default()) | |
110 | } |
|
119 | } | |
111 | } |
|
120 | } | |
112 |
|
121 | |||
113 | impl<'on_disk> ChildNodes<'on_disk> { |
|
122 | impl<'on_disk> ChildNodes<'on_disk> { | |
114 | pub(super) fn as_ref<'tree>( |
|
123 | pub(super) fn as_ref<'tree>( | |
115 | &'tree self, |
|
124 | &'tree self, | |
116 | ) -> ChildNodesRef<'tree, 'on_disk> { |
|
125 | ) -> ChildNodesRef<'tree, 'on_disk> { | |
117 | match self { |
|
126 | match self { | |
118 | ChildNodes::InMemory(nodes) => ChildNodesRef::InMemory(nodes), |
|
127 | ChildNodes::InMemory(nodes) => ChildNodesRef::InMemory(nodes), | |
119 | ChildNodes::OnDisk(nodes) => ChildNodesRef::OnDisk(nodes), |
|
128 | ChildNodes::OnDisk(nodes) => ChildNodesRef::OnDisk(nodes), | |
120 | } |
|
129 | } | |
121 | } |
|
130 | } | |
122 |
|
131 | |||
123 | pub(super) fn is_empty(&self) -> bool { |
|
132 | pub(super) fn is_empty(&self) -> bool { | |
124 | match self { |
|
133 | match self { | |
125 | ChildNodes::InMemory(nodes) => nodes.is_empty(), |
|
134 | ChildNodes::InMemory(nodes) => nodes.is_empty(), | |
126 | ChildNodes::OnDisk(nodes) => nodes.is_empty(), |
|
135 | ChildNodes::OnDisk(nodes) => nodes.is_empty(), | |
127 | } |
|
136 | } | |
128 | } |
|
137 | } | |
129 |
|
138 | |||
130 | fn make_mut( |
|
139 | fn make_mut( | |
131 | &mut self, |
|
140 | &mut self, | |
132 | on_disk: &'on_disk [u8], |
|
141 | on_disk: &'on_disk [u8], | |
133 | unreachable_bytes: &mut u32, |
|
142 | unreachable_bytes: &mut u32, | |
134 | ) -> Result< |
|
143 | ) -> Result< | |
135 | &mut FastHashMap<NodeKey<'on_disk>, Node<'on_disk>>, |
|
144 | &mut FastHashMap<NodeKey<'on_disk>, Node<'on_disk>>, | |
136 | DirstateV2ParseError, |
|
145 | DirstateV2ParseError, | |
137 | > { |
|
146 | > { | |
138 | match self { |
|
147 | match self { | |
139 | ChildNodes::InMemory(nodes) => Ok(nodes), |
|
148 | ChildNodes::InMemory(nodes) => Ok(nodes), | |
140 | ChildNodes::OnDisk(nodes) => { |
|
149 | ChildNodes::OnDisk(nodes) => { | |
141 | *unreachable_bytes += |
|
150 | *unreachable_bytes += | |
142 | std::mem::size_of_val::<[on_disk::Node]>(nodes) as u32; |
|
151 | std::mem::size_of_val::<[on_disk::Node]>(nodes) as u32; | |
143 | let nodes = nodes |
|
152 | let nodes = nodes | |
144 | .iter() |
|
153 | .iter() | |
145 | .map(|node| { |
|
154 | .map(|node| { | |
146 | Ok(( |
|
155 | Ok(( | |
147 | node.path(on_disk)?, |
|
156 | node.path(on_disk)?, | |
148 | node.to_in_memory_node(on_disk)?, |
|
157 | node.to_in_memory_node(on_disk)?, | |
149 | )) |
|
158 | )) | |
150 | }) |
|
159 | }) | |
151 | .collect::<Result<_, _>>()?; |
|
160 | .collect::<Result<_, _>>()?; | |
152 | *self = ChildNodes::InMemory(nodes); |
|
161 | *self = ChildNodes::InMemory(nodes); | |
153 | match self { |
|
162 | match self { | |
154 | ChildNodes::InMemory(nodes) => Ok(nodes), |
|
163 | ChildNodes::InMemory(nodes) => Ok(nodes), | |
155 | ChildNodes::OnDisk(_) => unreachable!(), |
|
164 | ChildNodes::OnDisk(_) => unreachable!(), | |
156 | } |
|
165 | } | |
157 | } |
|
166 | } | |
158 | } |
|
167 | } | |
159 | } |
|
168 | } | |
160 | } |
|
169 | } | |
161 |
|
170 | |||
162 | impl<'tree, 'on_disk> ChildNodesRef<'tree, 'on_disk> { |
|
171 | impl<'tree, 'on_disk> ChildNodesRef<'tree, 'on_disk> { | |
163 | pub(super) fn get( |
|
172 | pub(super) fn get( | |
164 | &self, |
|
173 | &self, | |
165 | base_name: &HgPath, |
|
174 | base_name: &HgPath, | |
166 | on_disk: &'on_disk [u8], |
|
175 | on_disk: &'on_disk [u8], | |
167 | ) -> Result<Option<NodeRef<'tree, 'on_disk>>, DirstateV2ParseError> { |
|
176 | ) -> Result<Option<NodeRef<'tree, 'on_disk>>, DirstateV2ParseError> { | |
168 | match self { |
|
177 | match self { | |
169 | ChildNodesRef::InMemory(nodes) => Ok(nodes |
|
178 | ChildNodesRef::InMemory(nodes) => Ok(nodes | |
170 | .get_key_value(base_name) |
|
179 | .get_key_value(base_name) | |
171 | .map(|(k, v)| NodeRef::InMemory(k, v))), |
|
180 | .map(|(k, v)| NodeRef::InMemory(k, v))), | |
172 | ChildNodesRef::OnDisk(nodes) => { |
|
181 | ChildNodesRef::OnDisk(nodes) => { | |
173 | let mut parse_result = Ok(()); |
|
182 | let mut parse_result = Ok(()); | |
174 | let search_result = nodes.binary_search_by(|node| { |
|
183 | let search_result = nodes.binary_search_by(|node| { | |
175 | match node.base_name(on_disk) { |
|
184 | match node.base_name(on_disk) { | |
176 | Ok(node_base_name) => node_base_name.cmp(base_name), |
|
185 | Ok(node_base_name) => node_base_name.cmp(base_name), | |
177 | Err(e) => { |
|
186 | Err(e) => { | |
178 | parse_result = Err(e); |
|
187 | parse_result = Err(e); | |
179 | // Dummy comparison result, `search_result` won’t |
|
188 | // Dummy comparison result, `search_result` won’t | |
180 | // be used since `parse_result` is an error |
|
189 | // be used since `parse_result` is an error | |
181 | std::cmp::Ordering::Equal |
|
190 | std::cmp::Ordering::Equal | |
182 | } |
|
191 | } | |
183 | } |
|
192 | } | |
184 | }); |
|
193 | }); | |
185 | parse_result.map(|()| { |
|
194 | parse_result.map(|()| { | |
186 | search_result.ok().map(|i| NodeRef::OnDisk(&nodes[i])) |
|
195 | search_result.ok().map(|i| NodeRef::OnDisk(&nodes[i])) | |
187 | }) |
|
196 | }) | |
188 | } |
|
197 | } | |
189 | } |
|
198 | } | |
190 | } |
|
199 | } | |
191 |
|
200 | |||
192 | /// Iterate in undefined order |
|
201 | /// Iterate in undefined order | |
193 | pub(super) fn iter( |
|
202 | pub(super) fn iter( | |
194 | &self, |
|
203 | &self, | |
195 | ) -> impl Iterator<Item = NodeRef<'tree, 'on_disk>> { |
|
204 | ) -> impl Iterator<Item = NodeRef<'tree, 'on_disk>> { | |
196 | match self { |
|
205 | match self { | |
197 | ChildNodesRef::InMemory(nodes) => itertools::Either::Left( |
|
206 | ChildNodesRef::InMemory(nodes) => itertools::Either::Left( | |
198 | nodes.iter().map(|(k, v)| NodeRef::InMemory(k, v)), |
|
207 | nodes.iter().map(|(k, v)| NodeRef::InMemory(k, v)), | |
199 | ), |
|
208 | ), | |
200 | ChildNodesRef::OnDisk(nodes) => { |
|
209 | ChildNodesRef::OnDisk(nodes) => { | |
201 | itertools::Either::Right(nodes.iter().map(NodeRef::OnDisk)) |
|
210 | itertools::Either::Right(nodes.iter().map(NodeRef::OnDisk)) | |
202 | } |
|
211 | } | |
203 | } |
|
212 | } | |
204 | } |
|
213 | } | |
205 |
|
214 | |||
206 | /// Iterate in parallel in undefined order |
|
215 | /// Iterate in parallel in undefined order | |
207 | pub(super) fn par_iter( |
|
216 | pub(super) fn par_iter( | |
208 | &self, |
|
217 | &self, | |
209 | ) -> impl rayon::iter::ParallelIterator<Item = NodeRef<'tree, 'on_disk>> |
|
218 | ) -> impl rayon::iter::ParallelIterator<Item = NodeRef<'tree, 'on_disk>> | |
210 | { |
|
219 | { | |
211 | use rayon::prelude::*; |
|
220 | use rayon::prelude::*; | |
212 | match self { |
|
221 | match self { | |
213 | ChildNodesRef::InMemory(nodes) => rayon::iter::Either::Left( |
|
222 | ChildNodesRef::InMemory(nodes) => rayon::iter::Either::Left( | |
214 | nodes.par_iter().map(|(k, v)| NodeRef::InMemory(k, v)), |
|
223 | nodes.par_iter().map(|(k, v)| NodeRef::InMemory(k, v)), | |
215 | ), |
|
224 | ), | |
216 | ChildNodesRef::OnDisk(nodes) => rayon::iter::Either::Right( |
|
225 | ChildNodesRef::OnDisk(nodes) => rayon::iter::Either::Right( | |
217 | nodes.par_iter().map(NodeRef::OnDisk), |
|
226 | nodes.par_iter().map(NodeRef::OnDisk), | |
218 | ), |
|
227 | ), | |
219 | } |
|
228 | } | |
220 | } |
|
229 | } | |
221 |
|
230 | |||
222 | pub(super) fn sorted(&self) -> Vec<NodeRef<'tree, 'on_disk>> { |
|
231 | pub(super) fn sorted(&self) -> Vec<NodeRef<'tree, 'on_disk>> { | |
223 | match self { |
|
232 | match self { | |
224 | ChildNodesRef::InMemory(nodes) => { |
|
233 | ChildNodesRef::InMemory(nodes) => { | |
225 | let mut vec: Vec<_> = nodes |
|
234 | let mut vec: Vec<_> = nodes | |
226 | .iter() |
|
235 | .iter() | |
227 | .map(|(k, v)| NodeRef::InMemory(k, v)) |
|
236 | .map(|(k, v)| NodeRef::InMemory(k, v)) | |
228 | .collect(); |
|
237 | .collect(); | |
229 | fn sort_key<'a>(node: &'a NodeRef) -> &'a HgPath { |
|
238 | fn sort_key<'a>(node: &'a NodeRef) -> &'a HgPath { | |
230 | match node { |
|
239 | match node { | |
231 | NodeRef::InMemory(path, _node) => path.base_name(), |
|
240 | NodeRef::InMemory(path, _node) => path.base_name(), | |
232 | NodeRef::OnDisk(_) => unreachable!(), |
|
241 | NodeRef::OnDisk(_) => unreachable!(), | |
233 | } |
|
242 | } | |
234 | } |
|
243 | } | |
235 | // `sort_unstable_by_key` doesn’t allow keys borrowing from the |
|
244 | // `sort_unstable_by_key` doesn’t allow keys borrowing from the | |
236 | // value: https://github.com/rust-lang/rust/issues/34162 |
|
245 | // value: https://github.com/rust-lang/rust/issues/34162 | |
237 | vec.sort_unstable_by(|a, b| sort_key(a).cmp(sort_key(b))); |
|
246 | vec.sort_unstable_by(|a, b| sort_key(a).cmp(sort_key(b))); | |
238 | vec |
|
247 | vec | |
239 | } |
|
248 | } | |
240 | ChildNodesRef::OnDisk(nodes) => { |
|
249 | ChildNodesRef::OnDisk(nodes) => { | |
241 | // Nodes on disk are already sorted |
|
250 | // Nodes on disk are already sorted | |
242 | nodes.iter().map(NodeRef::OnDisk).collect() |
|
251 | nodes.iter().map(NodeRef::OnDisk).collect() | |
243 | } |
|
252 | } | |
244 | } |
|
253 | } | |
245 | } |
|
254 | } | |
246 | } |
|
255 | } | |
247 |
|
256 | |||
248 | impl<'tree, 'on_disk> NodeRef<'tree, 'on_disk> { |
|
257 | impl<'tree, 'on_disk> NodeRef<'tree, 'on_disk> { | |
249 | pub(super) fn full_path( |
|
258 | pub(super) fn full_path( | |
250 | &self, |
|
259 | &self, | |
251 | on_disk: &'on_disk [u8], |
|
260 | on_disk: &'on_disk [u8], | |
252 | ) -> Result<&'tree HgPath, DirstateV2ParseError> { |
|
261 | ) -> Result<&'tree HgPath, DirstateV2ParseError> { | |
253 | match self { |
|
262 | match self { | |
254 | NodeRef::InMemory(path, _node) => Ok(path.full_path()), |
|
263 | NodeRef::InMemory(path, _node) => Ok(path.full_path()), | |
255 | NodeRef::OnDisk(node) => node.full_path(on_disk), |
|
264 | NodeRef::OnDisk(node) => node.full_path(on_disk), | |
256 | } |
|
265 | } | |
257 | } |
|
266 | } | |
258 |
|
267 | |||
259 | /// Returns a `BorrowedPath`, which can be turned into a `Cow<'on_disk, |
|
268 | /// Returns a `BorrowedPath`, which can be turned into a `Cow<'on_disk, | |
260 | /// HgPath>` detached from `'tree` |
|
269 | /// HgPath>` detached from `'tree` | |
261 | pub(super) fn full_path_borrowed( |
|
270 | pub(super) fn full_path_borrowed( | |
262 | &self, |
|
271 | &self, | |
263 | on_disk: &'on_disk [u8], |
|
272 | on_disk: &'on_disk [u8], | |
264 | ) -> Result<BorrowedPath<'tree, 'on_disk>, DirstateV2ParseError> { |
|
273 | ) -> Result<BorrowedPath<'tree, 'on_disk>, DirstateV2ParseError> { | |
265 | match self { |
|
274 | match self { | |
266 | NodeRef::InMemory(path, _node) => match path.full_path() { |
|
275 | NodeRef::InMemory(path, _node) => match path.full_path() { | |
267 | Cow::Borrowed(on_disk) => Ok(BorrowedPath::OnDisk(on_disk)), |
|
276 | Cow::Borrowed(on_disk) => Ok(BorrowedPath::OnDisk(on_disk)), | |
268 | Cow::Owned(in_memory) => Ok(BorrowedPath::InMemory(in_memory)), |
|
277 | Cow::Owned(in_memory) => Ok(BorrowedPath::InMemory(in_memory)), | |
269 | }, |
|
278 | }, | |
270 | NodeRef::OnDisk(node) => { |
|
279 | NodeRef::OnDisk(node) => { | |
271 | Ok(BorrowedPath::OnDisk(node.full_path(on_disk)?)) |
|
280 | Ok(BorrowedPath::OnDisk(node.full_path(on_disk)?)) | |
272 | } |
|
281 | } | |
273 | } |
|
282 | } | |
274 | } |
|
283 | } | |
275 |
|
284 | |||
276 | pub(super) fn base_name( |
|
285 | pub(super) fn base_name( | |
277 | &self, |
|
286 | &self, | |
278 | on_disk: &'on_disk [u8], |
|
287 | on_disk: &'on_disk [u8], | |
279 | ) -> Result<&'tree HgPath, DirstateV2ParseError> { |
|
288 | ) -> Result<&'tree HgPath, DirstateV2ParseError> { | |
280 | match self { |
|
289 | match self { | |
281 | NodeRef::InMemory(path, _node) => Ok(path.base_name()), |
|
290 | NodeRef::InMemory(path, _node) => Ok(path.base_name()), | |
282 | NodeRef::OnDisk(node) => node.base_name(on_disk), |
|
291 | NodeRef::OnDisk(node) => node.base_name(on_disk), | |
283 | } |
|
292 | } | |
284 | } |
|
293 | } | |
285 |
|
294 | |||
286 | pub(super) fn children( |
|
295 | pub(super) fn children( | |
287 | &self, |
|
296 | &self, | |
288 | on_disk: &'on_disk [u8], |
|
297 | on_disk: &'on_disk [u8], | |
289 | ) -> Result<ChildNodesRef<'tree, 'on_disk>, DirstateV2ParseError> { |
|
298 | ) -> Result<ChildNodesRef<'tree, 'on_disk>, DirstateV2ParseError> { | |
290 | match self { |
|
299 | match self { | |
291 | NodeRef::InMemory(_path, node) => Ok(node.children.as_ref()), |
|
300 | NodeRef::InMemory(_path, node) => Ok(node.children.as_ref()), | |
292 | NodeRef::OnDisk(node) => { |
|
301 | NodeRef::OnDisk(node) => { | |
293 | Ok(ChildNodesRef::OnDisk(node.children(on_disk)?)) |
|
302 | Ok(ChildNodesRef::OnDisk(node.children(on_disk)?)) | |
294 | } |
|
303 | } | |
295 | } |
|
304 | } | |
296 | } |
|
305 | } | |
297 |
|
306 | |||
298 | pub(super) fn has_copy_source(&self) -> bool { |
|
307 | pub(super) fn has_copy_source(&self) -> bool { | |
299 | match self { |
|
308 | match self { | |
300 | NodeRef::InMemory(_path, node) => node.copy_source.is_some(), |
|
309 | NodeRef::InMemory(_path, node) => node.copy_source.is_some(), | |
301 | NodeRef::OnDisk(node) => node.has_copy_source(), |
|
310 | NodeRef::OnDisk(node) => node.has_copy_source(), | |
302 | } |
|
311 | } | |
303 | } |
|
312 | } | |
304 |
|
313 | |||
305 | pub(super) fn copy_source( |
|
314 | pub(super) fn copy_source( | |
306 | &self, |
|
315 | &self, | |
307 | on_disk: &'on_disk [u8], |
|
316 | on_disk: &'on_disk [u8], | |
308 | ) -> Result<Option<&'tree HgPath>, DirstateV2ParseError> { |
|
317 | ) -> Result<Option<&'tree HgPath>, DirstateV2ParseError> { | |
309 | match self { |
|
318 | match self { | |
310 | NodeRef::InMemory(_path, node) => { |
|
319 | NodeRef::InMemory(_path, node) => { | |
311 | Ok(node.copy_source.as_ref().map(|s| &**s)) |
|
320 | Ok(node.copy_source.as_ref().map(|s| &**s)) | |
312 | } |
|
321 | } | |
313 | NodeRef::OnDisk(node) => node.copy_source(on_disk), |
|
322 | NodeRef::OnDisk(node) => node.copy_source(on_disk), | |
314 | } |
|
323 | } | |
315 | } |
|
324 | } | |
316 | /// Returns a `BorrowedPath`, which can be turned into a `Cow<'on_disk, |
|
325 | /// Returns a `BorrowedPath`, which can be turned into a `Cow<'on_disk, | |
317 | /// HgPath>` detached from `'tree` |
|
326 | /// HgPath>` detached from `'tree` | |
318 | pub(super) fn copy_source_borrowed( |
|
327 | pub(super) fn copy_source_borrowed( | |
319 | &self, |
|
328 | &self, | |
320 | on_disk: &'on_disk [u8], |
|
329 | on_disk: &'on_disk [u8], | |
321 | ) -> Result<Option<BorrowedPath<'tree, 'on_disk>>, DirstateV2ParseError> |
|
330 | ) -> Result<Option<BorrowedPath<'tree, 'on_disk>>, DirstateV2ParseError> | |
322 | { |
|
331 | { | |
323 | Ok(match self { |
|
332 | Ok(match self { | |
324 | NodeRef::InMemory(_path, node) => { |
|
333 | NodeRef::InMemory(_path, node) => { | |
325 | node.copy_source.as_ref().map(|source| match source { |
|
334 | node.copy_source.as_ref().map(|source| match source { | |
326 | Cow::Borrowed(on_disk) => BorrowedPath::OnDisk(on_disk), |
|
335 | Cow::Borrowed(on_disk) => BorrowedPath::OnDisk(on_disk), | |
327 | Cow::Owned(in_memory) => BorrowedPath::InMemory(in_memory), |
|
336 | Cow::Owned(in_memory) => BorrowedPath::InMemory(in_memory), | |
328 | }) |
|
337 | }) | |
329 | } |
|
338 | } | |
330 | NodeRef::OnDisk(node) => node |
|
339 | NodeRef::OnDisk(node) => node | |
331 | .copy_source(on_disk)? |
|
340 | .copy_source(on_disk)? | |
332 | .map(|source| BorrowedPath::OnDisk(source)), |
|
341 | .map(|source| BorrowedPath::OnDisk(source)), | |
333 | }) |
|
342 | }) | |
334 | } |
|
343 | } | |
335 |
|
344 | |||
336 | pub(super) fn entry( |
|
345 | pub(super) fn entry( | |
337 | &self, |
|
346 | &self, | |
338 | ) -> Result<Option<DirstateEntry>, DirstateV2ParseError> { |
|
347 | ) -> Result<Option<DirstateEntry>, DirstateV2ParseError> { | |
339 | match self { |
|
348 | match self { | |
340 | NodeRef::InMemory(_path, node) => { |
|
349 | NodeRef::InMemory(_path, node) => { | |
341 | Ok(node.data.as_entry().copied()) |
|
350 | Ok(node.data.as_entry().copied()) | |
342 | } |
|
351 | } | |
343 | NodeRef::OnDisk(node) => node.entry(), |
|
352 | NodeRef::OnDisk(node) => node.entry(), | |
344 | } |
|
353 | } | |
345 | } |
|
354 | } | |
346 |
|
355 | |||
347 | pub(super) fn state( |
|
356 | pub(super) fn state( | |
348 | &self, |
|
357 | &self, | |
349 | ) -> Result<Option<EntryState>, DirstateV2ParseError> { |
|
358 | ) -> Result<Option<EntryState>, DirstateV2ParseError> { | |
350 | Ok(self.entry()?.map(|e| e.state())) |
|
359 | Ok(self.entry()?.map(|e| e.state())) | |
351 | } |
|
360 | } | |
352 |
|
361 | |||
353 | pub(super) fn cached_directory_mtime( |
|
362 | pub(super) fn cached_directory_mtime( | |
354 | &self, |
|
363 | &self, | |
355 | ) -> Result<Option<TruncatedTimestamp>, DirstateV2ParseError> { |
|
364 | ) -> Result<Option<TruncatedTimestamp>, DirstateV2ParseError> { | |
356 | match self { |
|
365 | match self { | |
357 | NodeRef::InMemory(_path, node) => Ok(match node.data { |
|
366 | NodeRef::InMemory(_path, node) => Ok(match node.data { | |
358 | NodeData::CachedDirectory { mtime } => Some(mtime), |
|
367 | NodeData::CachedDirectory { mtime } => Some(mtime), | |
359 | _ => None, |
|
368 | _ => None, | |
360 | }), |
|
369 | }), | |
361 | NodeRef::OnDisk(node) => node.cached_directory_mtime(), |
|
370 | NodeRef::OnDisk(node) => node.cached_directory_mtime(), | |
362 | } |
|
371 | } | |
363 | } |
|
372 | } | |
364 |
|
373 | |||
365 | pub(super) fn descendants_with_entry_count(&self) -> u32 { |
|
374 | pub(super) fn descendants_with_entry_count(&self) -> u32 { | |
366 | match self { |
|
375 | match self { | |
367 | NodeRef::InMemory(_path, node) => { |
|
376 | NodeRef::InMemory(_path, node) => { | |
368 | node.descendants_with_entry_count |
|
377 | node.descendants_with_entry_count | |
369 | } |
|
378 | } | |
370 | NodeRef::OnDisk(node) => node.descendants_with_entry_count.get(), |
|
379 | NodeRef::OnDisk(node) => node.descendants_with_entry_count.get(), | |
371 | } |
|
380 | } | |
372 | } |
|
381 | } | |
373 |
|
382 | |||
374 | pub(super) fn tracked_descendants_count(&self) -> u32 { |
|
383 | pub(super) fn tracked_descendants_count(&self) -> u32 { | |
375 | match self { |
|
384 | match self { | |
376 | NodeRef::InMemory(_path, node) => node.tracked_descendants_count, |
|
385 | NodeRef::InMemory(_path, node) => node.tracked_descendants_count, | |
377 | NodeRef::OnDisk(node) => node.tracked_descendants_count.get(), |
|
386 | NodeRef::OnDisk(node) => node.tracked_descendants_count.get(), | |
378 | } |
|
387 | } | |
379 | } |
|
388 | } | |
380 | } |
|
389 | } | |
381 |
|
390 | |||
382 | /// Represents a file or a directory |
|
391 | /// Represents a file or a directory | |
383 | #[derive(Default)] |
|
392 | #[derive(Default)] | |
384 | pub(super) struct Node<'on_disk> { |
|
393 | pub(super) struct Node<'on_disk> { | |
385 | pub(super) data: NodeData, |
|
394 | pub(super) data: NodeData, | |
386 |
|
395 | |||
387 | pub(super) copy_source: Option<Cow<'on_disk, HgPath>>, |
|
396 | pub(super) copy_source: Option<Cow<'on_disk, HgPath>>, | |
388 |
|
397 | |||
389 | pub(super) children: ChildNodes<'on_disk>, |
|
398 | pub(super) children: ChildNodes<'on_disk>, | |
390 |
|
399 | |||
391 | /// How many (non-inclusive) descendants of this node have an entry. |
|
400 | /// How many (non-inclusive) descendants of this node have an entry. | |
392 | pub(super) descendants_with_entry_count: u32, |
|
401 | pub(super) descendants_with_entry_count: u32, | |
393 |
|
402 | |||
394 | /// How many (non-inclusive) descendants of this node have an entry whose |
|
403 | /// How many (non-inclusive) descendants of this node have an entry whose | |
395 | /// state is "tracked". |
|
404 | /// state is "tracked". | |
396 | pub(super) tracked_descendants_count: u32, |
|
405 | pub(super) tracked_descendants_count: u32, | |
397 | } |
|
406 | } | |
398 |
|
407 | |||
399 | pub(super) enum NodeData { |
|
408 | pub(super) enum NodeData { | |
400 | Entry(DirstateEntry), |
|
409 | Entry(DirstateEntry), | |
401 | CachedDirectory { mtime: TruncatedTimestamp }, |
|
410 | CachedDirectory { mtime: TruncatedTimestamp }, | |
402 | None, |
|
411 | None, | |
403 | } |
|
412 | } | |
404 |
|
413 | |||
405 | impl Default for NodeData { |
|
414 | impl Default for NodeData { | |
406 | fn default() -> Self { |
|
415 | fn default() -> Self { | |
407 | NodeData::None |
|
416 | NodeData::None | |
408 | } |
|
417 | } | |
409 | } |
|
418 | } | |
410 |
|
419 | |||
411 | impl NodeData { |
|
420 | impl NodeData { | |
412 | fn has_entry(&self) -> bool { |
|
421 | fn has_entry(&self) -> bool { | |
413 | match self { |
|
422 | match self { | |
414 | NodeData::Entry(_) => true, |
|
423 | NodeData::Entry(_) => true, | |
415 | _ => false, |
|
424 | _ => false, | |
416 | } |
|
425 | } | |
417 | } |
|
426 | } | |
418 |
|
427 | |||
419 | fn as_entry(&self) -> Option<&DirstateEntry> { |
|
428 | fn as_entry(&self) -> Option<&DirstateEntry> { | |
420 | match self { |
|
429 | match self { | |
421 | NodeData::Entry(entry) => Some(entry), |
|
430 | NodeData::Entry(entry) => Some(entry), | |
422 | _ => None, |
|
431 | _ => None, | |
423 | } |
|
432 | } | |
424 | } |
|
433 | } | |
425 | } |
|
434 | } | |
426 |
|
435 | |||
427 | impl<'on_disk> DirstateMap<'on_disk> { |
|
436 | impl<'on_disk> DirstateMap<'on_disk> { | |
428 | pub(super) fn empty(on_disk: &'on_disk [u8]) -> Self { |
|
437 | pub(super) fn empty(on_disk: &'on_disk [u8]) -> Self { | |
429 | Self { |
|
438 | Self { | |
430 | on_disk, |
|
439 | on_disk, | |
431 | root: ChildNodes::default(), |
|
440 | root: ChildNodes::default(), | |
432 | nodes_with_entry_count: 0, |
|
441 | nodes_with_entry_count: 0, | |
433 | nodes_with_copy_source_count: 0, |
|
442 | nodes_with_copy_source_count: 0, | |
434 | ignore_patterns_hash: [0; on_disk::IGNORE_PATTERNS_HASH_LEN], |
|
443 | ignore_patterns_hash: [0; on_disk::IGNORE_PATTERNS_HASH_LEN], | |
435 | unreachable_bytes: 0, |
|
444 | unreachable_bytes: 0, | |
436 | old_data_size: 0, |
|
445 | old_data_size: 0, | |
|
446 | dirstate_version: DirstateVersion::V1, | |||
437 | } |
|
447 | } | |
438 | } |
|
448 | } | |
439 |
|
449 | |||
440 | #[timed] |
|
450 | #[timed] | |
441 | pub fn new_v2( |
|
451 | pub fn new_v2( | |
442 | on_disk: &'on_disk [u8], |
|
452 | on_disk: &'on_disk [u8], | |
443 | data_size: usize, |
|
453 | data_size: usize, | |
444 | metadata: &[u8], |
|
454 | metadata: &[u8], | |
445 | ) -> Result<Self, DirstateError> { |
|
455 | ) -> Result<Self, DirstateError> { | |
446 | if let Some(data) = on_disk.get(..data_size) { |
|
456 | if let Some(data) = on_disk.get(..data_size) { | |
447 | Ok(on_disk::read(data, metadata)?) |
|
457 | Ok(on_disk::read(data, metadata)?) | |
448 | } else { |
|
458 | } else { | |
449 | Err(DirstateV2ParseError.into()) |
|
459 | Err(DirstateV2ParseError.into()) | |
450 | } |
|
460 | } | |
451 | } |
|
461 | } | |
452 |
|
462 | |||
453 | #[timed] |
|
463 | #[timed] | |
454 | pub fn new_v1( |
|
464 | pub fn new_v1( | |
455 | on_disk: &'on_disk [u8], |
|
465 | on_disk: &'on_disk [u8], | |
456 | ) -> Result<(Self, Option<DirstateParents>), DirstateError> { |
|
466 | ) -> Result<(Self, Option<DirstateParents>), DirstateError> { | |
457 | let mut map = Self::empty(on_disk); |
|
467 | let mut map = Self::empty(on_disk); | |
458 | if map.on_disk.is_empty() { |
|
468 | if map.on_disk.is_empty() { | |
459 | return Ok((map, None)); |
|
469 | return Ok((map, None)); | |
460 | } |
|
470 | } | |
461 |
|
471 | |||
462 | let parents = parse_dirstate_entries( |
|
472 | let parents = parse_dirstate_entries( | |
463 | map.on_disk, |
|
473 | map.on_disk, | |
464 | |path, entry, copy_source| { |
|
474 | |path, entry, copy_source| { | |
465 | let tracked = entry.state().is_tracked(); |
|
475 | let tracked = entry.state().is_tracked(); | |
466 | let node = Self::get_or_insert_node( |
|
476 | let node = Self::get_or_insert_node( | |
467 | map.on_disk, |
|
477 | map.on_disk, | |
468 | &mut map.unreachable_bytes, |
|
478 | &mut map.unreachable_bytes, | |
469 | &mut map.root, |
|
479 | &mut map.root, | |
470 | path, |
|
480 | path, | |
471 | WithBasename::to_cow_borrowed, |
|
481 | WithBasename::to_cow_borrowed, | |
472 | |ancestor| { |
|
482 | |ancestor| { | |
473 | if tracked { |
|
483 | if tracked { | |
474 | ancestor.tracked_descendants_count += 1 |
|
484 | ancestor.tracked_descendants_count += 1 | |
475 | } |
|
485 | } | |
476 | ancestor.descendants_with_entry_count += 1 |
|
486 | ancestor.descendants_with_entry_count += 1 | |
477 | }, |
|
487 | }, | |
478 | )?; |
|
488 | )?; | |
479 | assert!( |
|
489 | assert!( | |
480 | !node.data.has_entry(), |
|
490 | !node.data.has_entry(), | |
481 | "duplicate dirstate entry in read" |
|
491 | "duplicate dirstate entry in read" | |
482 | ); |
|
492 | ); | |
483 | assert!( |
|
493 | assert!( | |
484 | node.copy_source.is_none(), |
|
494 | node.copy_source.is_none(), | |
485 | "duplicate dirstate entry in read" |
|
495 | "duplicate dirstate entry in read" | |
486 | ); |
|
496 | ); | |
487 | node.data = NodeData::Entry(*entry); |
|
497 | node.data = NodeData::Entry(*entry); | |
488 | node.copy_source = copy_source.map(Cow::Borrowed); |
|
498 | node.copy_source = copy_source.map(Cow::Borrowed); | |
489 | map.nodes_with_entry_count += 1; |
|
499 | map.nodes_with_entry_count += 1; | |
490 | if copy_source.is_some() { |
|
500 | if copy_source.is_some() { | |
491 | map.nodes_with_copy_source_count += 1 |
|
501 | map.nodes_with_copy_source_count += 1 | |
492 | } |
|
502 | } | |
493 | Ok(()) |
|
503 | Ok(()) | |
494 | }, |
|
504 | }, | |
495 | )?; |
|
505 | )?; | |
496 | let parents = Some(parents.clone()); |
|
506 | let parents = Some(parents.clone()); | |
497 |
|
507 | |||
498 | Ok((map, parents)) |
|
508 | Ok((map, parents)) | |
499 | } |
|
509 | } | |
500 |
|
510 | |||
501 | /// Assuming dirstate-v2 format, returns whether the next write should |
|
511 | /// Assuming dirstate-v2 format, returns whether the next write should | |
502 | /// append to the existing data file that contains `self.on_disk` (true), |
|
512 | /// append to the existing data file that contains `self.on_disk` (true), | |
503 | /// or create a new data file from scratch (false). |
|
513 | /// or create a new data file from scratch (false). | |
504 | pub(super) fn write_should_append(&self) -> bool { |
|
514 | pub(super) fn write_should_append(&self) -> bool { | |
505 | let ratio = self.unreachable_bytes as f32 / self.on_disk.len() as f32; |
|
515 | let ratio = self.unreachable_bytes as f32 / self.on_disk.len() as f32; | |
506 | ratio < ACCEPTABLE_UNREACHABLE_BYTES_RATIO |
|
516 | ratio < ACCEPTABLE_UNREACHABLE_BYTES_RATIO | |
507 | } |
|
517 | } | |
508 |
|
518 | |||
509 | fn get_node<'tree>( |
|
519 | fn get_node<'tree>( | |
510 | &'tree self, |
|
520 | &'tree self, | |
511 | path: &HgPath, |
|
521 | path: &HgPath, | |
512 | ) -> Result<Option<NodeRef<'tree, 'on_disk>>, DirstateV2ParseError> { |
|
522 | ) -> Result<Option<NodeRef<'tree, 'on_disk>>, DirstateV2ParseError> { | |
513 | let mut children = self.root.as_ref(); |
|
523 | let mut children = self.root.as_ref(); | |
514 | let mut components = path.components(); |
|
524 | let mut components = path.components(); | |
515 | let mut component = |
|
525 | let mut component = | |
516 | components.next().expect("expected at least one components"); |
|
526 | components.next().expect("expected at least one components"); | |
517 | loop { |
|
527 | loop { | |
518 | if let Some(child) = children.get(component, self.on_disk)? { |
|
528 | if let Some(child) = children.get(component, self.on_disk)? { | |
519 | if let Some(next_component) = components.next() { |
|
529 | if let Some(next_component) = components.next() { | |
520 | component = next_component; |
|
530 | component = next_component; | |
521 | children = child.children(self.on_disk)?; |
|
531 | children = child.children(self.on_disk)?; | |
522 | } else { |
|
532 | } else { | |
523 | return Ok(Some(child)); |
|
533 | return Ok(Some(child)); | |
524 | } |
|
534 | } | |
525 | } else { |
|
535 | } else { | |
526 | return Ok(None); |
|
536 | return Ok(None); | |
527 | } |
|
537 | } | |
528 | } |
|
538 | } | |
529 | } |
|
539 | } | |
530 |
|
540 | |||
531 | /// Returns a mutable reference to the node at `path` if it exists |
|
541 | /// Returns a mutable reference to the node at `path` if it exists | |
532 | /// |
|
542 | /// | |
533 | /// This takes `root` instead of `&mut self` so that callers can mutate |
|
543 | /// This takes `root` instead of `&mut self` so that callers can mutate | |
534 | /// other fields while the returned borrow is still valid |
|
544 | /// other fields while the returned borrow is still valid | |
535 | fn get_node_mut<'tree>( |
|
545 | fn get_node_mut<'tree>( | |
536 | on_disk: &'on_disk [u8], |
|
546 | on_disk: &'on_disk [u8], | |
537 | unreachable_bytes: &mut u32, |
|
547 | unreachable_bytes: &mut u32, | |
538 | root: &'tree mut ChildNodes<'on_disk>, |
|
548 | root: &'tree mut ChildNodes<'on_disk>, | |
539 | path: &HgPath, |
|
549 | path: &HgPath, | |
540 | ) -> Result<Option<&'tree mut Node<'on_disk>>, DirstateV2ParseError> { |
|
550 | ) -> Result<Option<&'tree mut Node<'on_disk>>, DirstateV2ParseError> { | |
541 | let mut children = root; |
|
551 | let mut children = root; | |
542 | let mut components = path.components(); |
|
552 | let mut components = path.components(); | |
543 | let mut component = |
|
553 | let mut component = | |
544 | components.next().expect("expected at least one components"); |
|
554 | components.next().expect("expected at least one components"); | |
545 | loop { |
|
555 | loop { | |
546 | if let Some(child) = children |
|
556 | if let Some(child) = children | |
547 | .make_mut(on_disk, unreachable_bytes)? |
|
557 | .make_mut(on_disk, unreachable_bytes)? | |
548 | .get_mut(component) |
|
558 | .get_mut(component) | |
549 | { |
|
559 | { | |
550 | if let Some(next_component) = components.next() { |
|
560 | if let Some(next_component) = components.next() { | |
551 | component = next_component; |
|
561 | component = next_component; | |
552 | children = &mut child.children; |
|
562 | children = &mut child.children; | |
553 | } else { |
|
563 | } else { | |
554 | return Ok(Some(child)); |
|
564 | return Ok(Some(child)); | |
555 | } |
|
565 | } | |
556 | } else { |
|
566 | } else { | |
557 | return Ok(None); |
|
567 | return Ok(None); | |
558 | } |
|
568 | } | |
559 | } |
|
569 | } | |
560 | } |
|
570 | } | |
561 |
|
571 | |||
562 | pub(super) fn get_or_insert<'tree, 'path>( |
|
572 | pub(super) fn get_or_insert<'tree, 'path>( | |
563 | &'tree mut self, |
|
573 | &'tree mut self, | |
564 | path: &HgPath, |
|
574 | path: &HgPath, | |
565 | ) -> Result<&'tree mut Node<'on_disk>, DirstateV2ParseError> { |
|
575 | ) -> Result<&'tree mut Node<'on_disk>, DirstateV2ParseError> { | |
566 | Self::get_or_insert_node( |
|
576 | Self::get_or_insert_node( | |
567 | self.on_disk, |
|
577 | self.on_disk, | |
568 | &mut self.unreachable_bytes, |
|
578 | &mut self.unreachable_bytes, | |
569 | &mut self.root, |
|
579 | &mut self.root, | |
570 | path, |
|
580 | path, | |
571 | WithBasename::to_cow_owned, |
|
581 | WithBasename::to_cow_owned, | |
572 | |_| {}, |
|
582 | |_| {}, | |
573 | ) |
|
583 | ) | |
574 | } |
|
584 | } | |
575 |
|
585 | |||
576 | fn get_or_insert_node<'tree, 'path>( |
|
586 | fn get_or_insert_node<'tree, 'path>( | |
577 | on_disk: &'on_disk [u8], |
|
587 | on_disk: &'on_disk [u8], | |
578 | unreachable_bytes: &mut u32, |
|
588 | unreachable_bytes: &mut u32, | |
579 | root: &'tree mut ChildNodes<'on_disk>, |
|
589 | root: &'tree mut ChildNodes<'on_disk>, | |
580 | path: &'path HgPath, |
|
590 | path: &'path HgPath, | |
581 | to_cow: impl Fn( |
|
591 | to_cow: impl Fn( | |
582 | WithBasename<&'path HgPath>, |
|
592 | WithBasename<&'path HgPath>, | |
583 | ) -> WithBasename<Cow<'on_disk, HgPath>>, |
|
593 | ) -> WithBasename<Cow<'on_disk, HgPath>>, | |
584 | mut each_ancestor: impl FnMut(&mut Node), |
|
594 | mut each_ancestor: impl FnMut(&mut Node), | |
585 | ) -> Result<&'tree mut Node<'on_disk>, DirstateV2ParseError> { |
|
595 | ) -> Result<&'tree mut Node<'on_disk>, DirstateV2ParseError> { | |
586 | let mut child_nodes = root; |
|
596 | let mut child_nodes = root; | |
587 | let mut inclusive_ancestor_paths = |
|
597 | let mut inclusive_ancestor_paths = | |
588 | WithBasename::inclusive_ancestors_of(path); |
|
598 | WithBasename::inclusive_ancestors_of(path); | |
589 | let mut ancestor_path = inclusive_ancestor_paths |
|
599 | let mut ancestor_path = inclusive_ancestor_paths | |
590 | .next() |
|
600 | .next() | |
591 | .expect("expected at least one inclusive ancestor"); |
|
601 | .expect("expected at least one inclusive ancestor"); | |
592 | loop { |
|
602 | loop { | |
593 | // TODO: can we avoid allocating an owned key in cases where the |
|
603 | // TODO: can we avoid allocating an owned key in cases where the | |
594 | // map already contains that key, without introducing double |
|
604 | // map already contains that key, without introducing double | |
595 | // lookup? |
|
605 | // lookup? | |
596 | let child_node = child_nodes |
|
606 | let child_node = child_nodes | |
597 | .make_mut(on_disk, unreachable_bytes)? |
|
607 | .make_mut(on_disk, unreachable_bytes)? | |
598 | .entry(to_cow(ancestor_path)) |
|
608 | .entry(to_cow(ancestor_path)) | |
599 | .or_default(); |
|
609 | .or_default(); | |
600 | if let Some(next) = inclusive_ancestor_paths.next() { |
|
610 | if let Some(next) = inclusive_ancestor_paths.next() { | |
601 | each_ancestor(child_node); |
|
611 | each_ancestor(child_node); | |
602 | ancestor_path = next; |
|
612 | ancestor_path = next; | |
603 | child_nodes = &mut child_node.children; |
|
613 | child_nodes = &mut child_node.children; | |
604 | } else { |
|
614 | } else { | |
605 | return Ok(child_node); |
|
615 | return Ok(child_node); | |
606 | } |
|
616 | } | |
607 | } |
|
617 | } | |
608 | } |
|
618 | } | |
609 |
|
619 | |||
610 | fn add_or_remove_file( |
|
620 | fn add_or_remove_file( | |
611 | &mut self, |
|
621 | &mut self, | |
612 | path: &HgPath, |
|
622 | path: &HgPath, | |
613 | old_state: Option<EntryState>, |
|
623 | old_state: Option<EntryState>, | |
614 | new_entry: DirstateEntry, |
|
624 | new_entry: DirstateEntry, | |
615 | ) -> Result<(), DirstateV2ParseError> { |
|
625 | ) -> Result<(), DirstateV2ParseError> { | |
616 | let had_entry = old_state.is_some(); |
|
626 | let had_entry = old_state.is_some(); | |
617 | let was_tracked = old_state.map_or(false, |s| s.is_tracked()); |
|
627 | let was_tracked = old_state.map_or(false, |s| s.is_tracked()); | |
618 | let tracked_count_increment = |
|
628 | let tracked_count_increment = | |
619 | match (was_tracked, new_entry.state().is_tracked()) { |
|
629 | match (was_tracked, new_entry.state().is_tracked()) { | |
620 | (false, true) => 1, |
|
630 | (false, true) => 1, | |
621 | (true, false) => -1, |
|
631 | (true, false) => -1, | |
622 | _ => 0, |
|
632 | _ => 0, | |
623 | }; |
|
633 | }; | |
624 |
|
634 | |||
625 | let node = Self::get_or_insert_node( |
|
635 | let node = Self::get_or_insert_node( | |
626 | self.on_disk, |
|
636 | self.on_disk, | |
627 | &mut self.unreachable_bytes, |
|
637 | &mut self.unreachable_bytes, | |
628 | &mut self.root, |
|
638 | &mut self.root, | |
629 | path, |
|
639 | path, | |
630 | WithBasename::to_cow_owned, |
|
640 | WithBasename::to_cow_owned, | |
631 | |ancestor| { |
|
641 | |ancestor| { | |
632 | if !had_entry { |
|
642 | if !had_entry { | |
633 | ancestor.descendants_with_entry_count += 1; |
|
643 | ancestor.descendants_with_entry_count += 1; | |
634 | } |
|
644 | } | |
635 |
|
645 | |||
636 | // We can’t use `+= increment` because the counter is unsigned, |
|
646 | // We can’t use `+= increment` because the counter is unsigned, | |
637 | // and we want debug builds to detect accidental underflow |
|
647 | // and we want debug builds to detect accidental underflow | |
638 | // through zero |
|
648 | // through zero | |
639 | match tracked_count_increment { |
|
649 | match tracked_count_increment { | |
640 | 1 => ancestor.tracked_descendants_count += 1, |
|
650 | 1 => ancestor.tracked_descendants_count += 1, | |
641 | -1 => ancestor.tracked_descendants_count -= 1, |
|
651 | -1 => ancestor.tracked_descendants_count -= 1, | |
642 | _ => {} |
|
652 | _ => {} | |
643 | } |
|
653 | } | |
644 | }, |
|
654 | }, | |
645 | )?; |
|
655 | )?; | |
646 | if !had_entry { |
|
656 | if !had_entry { | |
647 | self.nodes_with_entry_count += 1 |
|
657 | self.nodes_with_entry_count += 1 | |
648 | } |
|
658 | } | |
649 | node.data = NodeData::Entry(new_entry); |
|
659 | node.data = NodeData::Entry(new_entry); | |
650 | Ok(()) |
|
660 | Ok(()) | |
651 | } |
|
661 | } | |
652 |
|
662 | |||
653 | fn iter_nodes<'tree>( |
|
663 | fn iter_nodes<'tree>( | |
654 | &'tree self, |
|
664 | &'tree self, | |
655 | ) -> impl Iterator< |
|
665 | ) -> impl Iterator< | |
656 | Item = Result<NodeRef<'tree, 'on_disk>, DirstateV2ParseError>, |
|
666 | Item = Result<NodeRef<'tree, 'on_disk>, DirstateV2ParseError>, | |
657 | > + 'tree { |
|
667 | > + 'tree { | |
658 | // Depth first tree traversal. |
|
668 | // Depth first tree traversal. | |
659 | // |
|
669 | // | |
660 | // If we could afford internal iteration and recursion, |
|
670 | // If we could afford internal iteration and recursion, | |
661 | // this would look like: |
|
671 | // this would look like: | |
662 | // |
|
672 | // | |
663 | // ``` |
|
673 | // ``` | |
664 | // fn traverse_children( |
|
674 | // fn traverse_children( | |
665 | // children: &ChildNodes, |
|
675 | // children: &ChildNodes, | |
666 | // each: &mut impl FnMut(&Node), |
|
676 | // each: &mut impl FnMut(&Node), | |
667 | // ) { |
|
677 | // ) { | |
668 | // for child in children.values() { |
|
678 | // for child in children.values() { | |
669 | // traverse_children(&child.children, each); |
|
679 | // traverse_children(&child.children, each); | |
670 | // each(child); |
|
680 | // each(child); | |
671 | // } |
|
681 | // } | |
672 | // } |
|
682 | // } | |
673 | // ``` |
|
683 | // ``` | |
674 | // |
|
684 | // | |
675 | // However we want an external iterator and therefore can’t use the |
|
685 | // However we want an external iterator and therefore can’t use the | |
676 | // call stack. Use an explicit stack instead: |
|
686 | // call stack. Use an explicit stack instead: | |
677 | let mut stack = Vec::new(); |
|
687 | let mut stack = Vec::new(); | |
678 | let mut iter = self.root.as_ref().iter(); |
|
688 | let mut iter = self.root.as_ref().iter(); | |
679 | std::iter::from_fn(move || { |
|
689 | std::iter::from_fn(move || { | |
680 | while let Some(child_node) = iter.next() { |
|
690 | while let Some(child_node) = iter.next() { | |
681 | let children = match child_node.children(self.on_disk) { |
|
691 | let children = match child_node.children(self.on_disk) { | |
682 | Ok(children) => children, |
|
692 | Ok(children) => children, | |
683 | Err(error) => return Some(Err(error)), |
|
693 | Err(error) => return Some(Err(error)), | |
684 | }; |
|
694 | }; | |
685 | // Pseudo-recursion |
|
695 | // Pseudo-recursion | |
686 | let new_iter = children.iter(); |
|
696 | let new_iter = children.iter(); | |
687 | let old_iter = std::mem::replace(&mut iter, new_iter); |
|
697 | let old_iter = std::mem::replace(&mut iter, new_iter); | |
688 | stack.push((child_node, old_iter)); |
|
698 | stack.push((child_node, old_iter)); | |
689 | } |
|
699 | } | |
690 | // Found the end of a `children.iter()` iterator. |
|
700 | // Found the end of a `children.iter()` iterator. | |
691 | if let Some((child_node, next_iter)) = stack.pop() { |
|
701 | if let Some((child_node, next_iter)) = stack.pop() { | |
692 | // "Return" from pseudo-recursion by restoring state from the |
|
702 | // "Return" from pseudo-recursion by restoring state from the | |
693 | // explicit stack |
|
703 | // explicit stack | |
694 | iter = next_iter; |
|
704 | iter = next_iter; | |
695 |
|
705 | |||
696 | Some(Ok(child_node)) |
|
706 | Some(Ok(child_node)) | |
697 | } else { |
|
707 | } else { | |
698 | // Reached the bottom of the stack, we’re done |
|
708 | // Reached the bottom of the stack, we’re done | |
699 | None |
|
709 | None | |
700 | } |
|
710 | } | |
701 | }) |
|
711 | }) | |
702 | } |
|
712 | } | |
703 |
|
713 | |||
704 | fn count_dropped_path(unreachable_bytes: &mut u32, path: &Cow<HgPath>) { |
|
714 | fn count_dropped_path(unreachable_bytes: &mut u32, path: &Cow<HgPath>) { | |
705 | if let Cow::Borrowed(path) = path { |
|
715 | if let Cow::Borrowed(path) = path { | |
706 | *unreachable_bytes += path.len() as u32 |
|
716 | *unreachable_bytes += path.len() as u32 | |
707 | } |
|
717 | } | |
708 | } |
|
718 | } | |
709 | } |
|
719 | } | |
710 |
|
720 | |||
711 | /// Like `Iterator::filter_map`, but over a fallible iterator of `Result`s. |
|
721 | /// Like `Iterator::filter_map`, but over a fallible iterator of `Result`s. | |
712 | /// |
|
722 | /// | |
713 | /// The callback is only called for incoming `Ok` values. Errors are passed |
|
723 | /// The callback is only called for incoming `Ok` values. Errors are passed | |
714 | /// through as-is. In order to let it use the `?` operator the callback is |
|
724 | /// through as-is. In order to let it use the `?` operator the callback is | |
715 | /// expected to return a `Result` of `Option`, instead of an `Option` of |
|
725 | /// expected to return a `Result` of `Option`, instead of an `Option` of | |
716 | /// `Result`. |
|
726 | /// `Result`. | |
717 | fn filter_map_results<'a, I, F, A, B, E>( |
|
727 | fn filter_map_results<'a, I, F, A, B, E>( | |
718 | iter: I, |
|
728 | iter: I, | |
719 | f: F, |
|
729 | f: F, | |
720 | ) -> impl Iterator<Item = Result<B, E>> + 'a |
|
730 | ) -> impl Iterator<Item = Result<B, E>> + 'a | |
721 | where |
|
731 | where | |
722 | I: Iterator<Item = Result<A, E>> + 'a, |
|
732 | I: Iterator<Item = Result<A, E>> + 'a, | |
723 | F: Fn(A) -> Result<Option<B>, E> + 'a, |
|
733 | F: Fn(A) -> Result<Option<B>, E> + 'a, | |
724 | { |
|
734 | { | |
725 | iter.filter_map(move |result| match result { |
|
735 | iter.filter_map(move |result| match result { | |
726 | Ok(node) => f(node).transpose(), |
|
736 | Ok(node) => f(node).transpose(), | |
727 | Err(e) => Some(Err(e)), |
|
737 | Err(e) => Some(Err(e)), | |
728 | }) |
|
738 | }) | |
729 | } |
|
739 | } | |
730 |
|
740 | |||
731 | impl OwningDirstateMap { |
|
741 | impl OwningDirstateMap { | |
732 | pub fn clear(&mut self) { |
|
742 | pub fn clear(&mut self) { | |
733 | self.with_dmap_mut(|map| { |
|
743 | self.with_dmap_mut(|map| { | |
734 | map.root = Default::default(); |
|
744 | map.root = Default::default(); | |
735 | map.nodes_with_entry_count = 0; |
|
745 | map.nodes_with_entry_count = 0; | |
736 | map.nodes_with_copy_source_count = 0; |
|
746 | map.nodes_with_copy_source_count = 0; | |
737 | }); |
|
747 | }); | |
738 | } |
|
748 | } | |
739 |
|
749 | |||
740 | pub fn set_entry( |
|
750 | pub fn set_entry( | |
741 | &mut self, |
|
751 | &mut self, | |
742 | filename: &HgPath, |
|
752 | filename: &HgPath, | |
743 | entry: DirstateEntry, |
|
753 | entry: DirstateEntry, | |
744 | ) -> Result<(), DirstateV2ParseError> { |
|
754 | ) -> Result<(), DirstateV2ParseError> { | |
745 | self.with_dmap_mut(|map| { |
|
755 | self.with_dmap_mut(|map| { | |
746 | map.get_or_insert(&filename)?.data = NodeData::Entry(entry); |
|
756 | map.get_or_insert(&filename)?.data = NodeData::Entry(entry); | |
747 | Ok(()) |
|
757 | Ok(()) | |
748 | }) |
|
758 | }) | |
749 | } |
|
759 | } | |
750 |
|
760 | |||
751 | pub fn add_file( |
|
761 | pub fn add_file( | |
752 | &mut self, |
|
762 | &mut self, | |
753 | filename: &HgPath, |
|
763 | filename: &HgPath, | |
754 | entry: DirstateEntry, |
|
764 | entry: DirstateEntry, | |
755 | ) -> Result<(), DirstateError> { |
|
765 | ) -> Result<(), DirstateError> { | |
756 | let old_state = self.get(filename)?.map(|e| e.state()); |
|
766 | let old_state = self.get(filename)?.map(|e| e.state()); | |
757 | self.with_dmap_mut(|map| { |
|
767 | self.with_dmap_mut(|map| { | |
758 | Ok(map.add_or_remove_file(filename, old_state, entry)?) |
|
768 | Ok(map.add_or_remove_file(filename, old_state, entry)?) | |
759 | }) |
|
769 | }) | |
760 | } |
|
770 | } | |
761 |
|
771 | |||
762 | pub fn remove_file( |
|
772 | pub fn remove_file( | |
763 | &mut self, |
|
773 | &mut self, | |
764 | filename: &HgPath, |
|
774 | filename: &HgPath, | |
765 | in_merge: bool, |
|
775 | in_merge: bool, | |
766 | ) -> Result<(), DirstateError> { |
|
776 | ) -> Result<(), DirstateError> { | |
767 | let old_entry_opt = self.get(filename)?; |
|
777 | let old_entry_opt = self.get(filename)?; | |
768 | let old_state = old_entry_opt.map(|e| e.state()); |
|
778 | let old_state = old_entry_opt.map(|e| e.state()); | |
769 | let mut size = 0; |
|
779 | let mut size = 0; | |
770 | if in_merge { |
|
780 | if in_merge { | |
771 | // XXX we should not be able to have 'm' state and 'FROM_P2' if not |
|
781 | // XXX we should not be able to have 'm' state and 'FROM_P2' if not | |
772 | // during a merge. So I (marmoute) am not sure we need the |
|
782 | // during a merge. So I (marmoute) am not sure we need the | |
773 | // conditionnal at all. Adding double checking this with assert |
|
783 | // conditionnal at all. Adding double checking this with assert | |
774 | // would be nice. |
|
784 | // would be nice. | |
775 | if let Some(old_entry) = old_entry_opt { |
|
785 | if let Some(old_entry) = old_entry_opt { | |
776 | // backup the previous state |
|
786 | // backup the previous state | |
777 | if old_entry.state() == EntryState::Merged { |
|
787 | if old_entry.state() == EntryState::Merged { | |
778 | size = SIZE_NON_NORMAL; |
|
788 | size = SIZE_NON_NORMAL; | |
779 | } else if old_entry.state() == EntryState::Normal |
|
789 | } else if old_entry.state() == EntryState::Normal | |
780 | && old_entry.size() == SIZE_FROM_OTHER_PARENT |
|
790 | && old_entry.size() == SIZE_FROM_OTHER_PARENT | |
781 | { |
|
791 | { | |
782 | // other parent |
|
792 | // other parent | |
783 | size = SIZE_FROM_OTHER_PARENT; |
|
793 | size = SIZE_FROM_OTHER_PARENT; | |
784 | } |
|
794 | } | |
785 | } |
|
795 | } | |
786 | } |
|
796 | } | |
787 | if size == 0 { |
|
797 | if size == 0 { | |
788 | self.copy_map_remove(filename)?; |
|
798 | self.copy_map_remove(filename)?; | |
789 | } |
|
799 | } | |
790 | self.with_dmap_mut(|map| { |
|
800 | self.with_dmap_mut(|map| { | |
791 | let entry = DirstateEntry::new_removed(size); |
|
801 | let entry = DirstateEntry::new_removed(size); | |
792 | Ok(map.add_or_remove_file(filename, old_state, entry)?) |
|
802 | Ok(map.add_or_remove_file(filename, old_state, entry)?) | |
793 | }) |
|
803 | }) | |
794 | } |
|
804 | } | |
795 |
|
805 | |||
796 | pub fn drop_entry_and_copy_source( |
|
806 | pub fn drop_entry_and_copy_source( | |
797 | &mut self, |
|
807 | &mut self, | |
798 | filename: &HgPath, |
|
808 | filename: &HgPath, | |
799 | ) -> Result<(), DirstateError> { |
|
809 | ) -> Result<(), DirstateError> { | |
800 | let was_tracked = self |
|
810 | let was_tracked = self | |
801 | .get(filename)? |
|
811 | .get(filename)? | |
802 | .map_or(false, |e| e.state().is_tracked()); |
|
812 | .map_or(false, |e| e.state().is_tracked()); | |
803 | struct Dropped { |
|
813 | struct Dropped { | |
804 | was_tracked: bool, |
|
814 | was_tracked: bool, | |
805 | had_entry: bool, |
|
815 | had_entry: bool, | |
806 | had_copy_source: bool, |
|
816 | had_copy_source: bool, | |
807 | } |
|
817 | } | |
808 |
|
818 | |||
809 | /// If this returns `Ok(Some((dropped, removed)))`, then |
|
819 | /// If this returns `Ok(Some((dropped, removed)))`, then | |
810 | /// |
|
820 | /// | |
811 | /// * `dropped` is about the leaf node that was at `filename` |
|
821 | /// * `dropped` is about the leaf node that was at `filename` | |
812 | /// * `removed` is whether this particular level of recursion just |
|
822 | /// * `removed` is whether this particular level of recursion just | |
813 | /// removed a node in `nodes`. |
|
823 | /// removed a node in `nodes`. | |
814 | fn recur<'on_disk>( |
|
824 | fn recur<'on_disk>( | |
815 | on_disk: &'on_disk [u8], |
|
825 | on_disk: &'on_disk [u8], | |
816 | unreachable_bytes: &mut u32, |
|
826 | unreachable_bytes: &mut u32, | |
817 | nodes: &mut ChildNodes<'on_disk>, |
|
827 | nodes: &mut ChildNodes<'on_disk>, | |
818 | path: &HgPath, |
|
828 | path: &HgPath, | |
819 | ) -> Result<Option<(Dropped, bool)>, DirstateV2ParseError> { |
|
829 | ) -> Result<Option<(Dropped, bool)>, DirstateV2ParseError> { | |
820 | let (first_path_component, rest_of_path) = |
|
830 | let (first_path_component, rest_of_path) = | |
821 | path.split_first_component(); |
|
831 | path.split_first_component(); | |
822 | let nodes = nodes.make_mut(on_disk, unreachable_bytes)?; |
|
832 | let nodes = nodes.make_mut(on_disk, unreachable_bytes)?; | |
823 | let node = if let Some(node) = nodes.get_mut(first_path_component) |
|
833 | let node = if let Some(node) = nodes.get_mut(first_path_component) | |
824 | { |
|
834 | { | |
825 | node |
|
835 | node | |
826 | } else { |
|
836 | } else { | |
827 | return Ok(None); |
|
837 | return Ok(None); | |
828 | }; |
|
838 | }; | |
829 | let dropped; |
|
839 | let dropped; | |
830 | if let Some(rest) = rest_of_path { |
|
840 | if let Some(rest) = rest_of_path { | |
831 | if let Some((d, removed)) = recur( |
|
841 | if let Some((d, removed)) = recur( | |
832 | on_disk, |
|
842 | on_disk, | |
833 | unreachable_bytes, |
|
843 | unreachable_bytes, | |
834 | &mut node.children, |
|
844 | &mut node.children, | |
835 | rest, |
|
845 | rest, | |
836 | )? { |
|
846 | )? { | |
837 | dropped = d; |
|
847 | dropped = d; | |
838 | if dropped.had_entry { |
|
848 | if dropped.had_entry { | |
839 | node.descendants_with_entry_count = node |
|
849 | node.descendants_with_entry_count = node | |
840 | .descendants_with_entry_count |
|
850 | .descendants_with_entry_count | |
841 | .checked_sub(1) |
|
851 | .checked_sub(1) | |
842 | .expect( |
|
852 | .expect( | |
843 | "descendants_with_entry_count should be >= 0", |
|
853 | "descendants_with_entry_count should be >= 0", | |
844 | ); |
|
854 | ); | |
845 | } |
|
855 | } | |
846 | if dropped.was_tracked { |
|
856 | if dropped.was_tracked { | |
847 | node.tracked_descendants_count = node |
|
857 | node.tracked_descendants_count = node | |
848 | .tracked_descendants_count |
|
858 | .tracked_descendants_count | |
849 | .checked_sub(1) |
|
859 | .checked_sub(1) | |
850 | .expect( |
|
860 | .expect( | |
851 | "tracked_descendants_count should be >= 0", |
|
861 | "tracked_descendants_count should be >= 0", | |
852 | ); |
|
862 | ); | |
853 | } |
|
863 | } | |
854 |
|
864 | |||
855 | // Directory caches must be invalidated when removing a |
|
865 | // Directory caches must be invalidated when removing a | |
856 | // child node |
|
866 | // child node | |
857 | if removed { |
|
867 | if removed { | |
858 | if let NodeData::CachedDirectory { .. } = &node.data { |
|
868 | if let NodeData::CachedDirectory { .. } = &node.data { | |
859 | node.data = NodeData::None |
|
869 | node.data = NodeData::None | |
860 | } |
|
870 | } | |
861 | } |
|
871 | } | |
862 | } else { |
|
872 | } else { | |
863 | return Ok(None); |
|
873 | return Ok(None); | |
864 | } |
|
874 | } | |
865 | } else { |
|
875 | } else { | |
866 | let entry = node.data.as_entry(); |
|
876 | let entry = node.data.as_entry(); | |
867 | let was_tracked = entry.map_or(false, |entry| entry.tracked()); |
|
877 | let was_tracked = entry.map_or(false, |entry| entry.tracked()); | |
868 | let had_entry = entry.is_some(); |
|
878 | let had_entry = entry.is_some(); | |
869 | if had_entry { |
|
879 | if had_entry { | |
870 | node.data = NodeData::None |
|
880 | node.data = NodeData::None | |
871 | } |
|
881 | } | |
872 | let mut had_copy_source = false; |
|
882 | let mut had_copy_source = false; | |
873 | if let Some(source) = &node.copy_source { |
|
883 | if let Some(source) = &node.copy_source { | |
874 | DirstateMap::count_dropped_path(unreachable_bytes, source); |
|
884 | DirstateMap::count_dropped_path(unreachable_bytes, source); | |
875 | had_copy_source = true; |
|
885 | had_copy_source = true; | |
876 | node.copy_source = None |
|
886 | node.copy_source = None | |
877 | } |
|
887 | } | |
878 | dropped = Dropped { |
|
888 | dropped = Dropped { | |
879 | was_tracked, |
|
889 | was_tracked, | |
880 | had_entry, |
|
890 | had_entry, | |
881 | had_copy_source, |
|
891 | had_copy_source, | |
882 | }; |
|
892 | }; | |
883 | } |
|
893 | } | |
884 | // After recursion, for both leaf (rest_of_path is None) nodes and |
|
894 | // After recursion, for both leaf (rest_of_path is None) nodes and | |
885 | // parent nodes, remove a node if it just became empty. |
|
895 | // parent nodes, remove a node if it just became empty. | |
886 | let remove = !node.data.has_entry() |
|
896 | let remove = !node.data.has_entry() | |
887 | && node.copy_source.is_none() |
|
897 | && node.copy_source.is_none() | |
888 | && node.children.is_empty(); |
|
898 | && node.children.is_empty(); | |
889 | if remove { |
|
899 | if remove { | |
890 | let (key, _) = |
|
900 | let (key, _) = | |
891 | nodes.remove_entry(first_path_component).unwrap(); |
|
901 | nodes.remove_entry(first_path_component).unwrap(); | |
892 | DirstateMap::count_dropped_path( |
|
902 | DirstateMap::count_dropped_path( | |
893 | unreachable_bytes, |
|
903 | unreachable_bytes, | |
894 | key.full_path(), |
|
904 | key.full_path(), | |
895 | ) |
|
905 | ) | |
896 | } |
|
906 | } | |
897 | Ok(Some((dropped, remove))) |
|
907 | Ok(Some((dropped, remove))) | |
898 | } |
|
908 | } | |
899 |
|
909 | |||
900 | self.with_dmap_mut(|map| { |
|
910 | self.with_dmap_mut(|map| { | |
901 | if let Some((dropped, _removed)) = recur( |
|
911 | if let Some((dropped, _removed)) = recur( | |
902 | map.on_disk, |
|
912 | map.on_disk, | |
903 | &mut map.unreachable_bytes, |
|
913 | &mut map.unreachable_bytes, | |
904 | &mut map.root, |
|
914 | &mut map.root, | |
905 | filename, |
|
915 | filename, | |
906 | )? { |
|
916 | )? { | |
907 | if dropped.had_entry { |
|
917 | if dropped.had_entry { | |
908 | map.nodes_with_entry_count = map |
|
918 | map.nodes_with_entry_count = map | |
909 | .nodes_with_entry_count |
|
919 | .nodes_with_entry_count | |
910 | .checked_sub(1) |
|
920 | .checked_sub(1) | |
911 | .expect("nodes_with_entry_count should be >= 0"); |
|
921 | .expect("nodes_with_entry_count should be >= 0"); | |
912 | } |
|
922 | } | |
913 | if dropped.had_copy_source { |
|
923 | if dropped.had_copy_source { | |
914 | map.nodes_with_copy_source_count = map |
|
924 | map.nodes_with_copy_source_count = map | |
915 | .nodes_with_copy_source_count |
|
925 | .nodes_with_copy_source_count | |
916 | .checked_sub(1) |
|
926 | .checked_sub(1) | |
917 | .expect("nodes_with_copy_source_count should be >= 0"); |
|
927 | .expect("nodes_with_copy_source_count should be >= 0"); | |
918 | } |
|
928 | } | |
919 | } else { |
|
929 | } else { | |
920 | debug_assert!(!was_tracked); |
|
930 | debug_assert!(!was_tracked); | |
921 | } |
|
931 | } | |
922 | Ok(()) |
|
932 | Ok(()) | |
923 | }) |
|
933 | }) | |
924 | } |
|
934 | } | |
925 |
|
935 | |||
926 | pub fn has_tracked_dir( |
|
936 | pub fn has_tracked_dir( | |
927 | &mut self, |
|
937 | &mut self, | |
928 | directory: &HgPath, |
|
938 | directory: &HgPath, | |
929 | ) -> Result<bool, DirstateError> { |
|
939 | ) -> Result<bool, DirstateError> { | |
930 | self.with_dmap_mut(|map| { |
|
940 | self.with_dmap_mut(|map| { | |
931 | if let Some(node) = map.get_node(directory)? { |
|
941 | if let Some(node) = map.get_node(directory)? { | |
932 | // A node without a `DirstateEntry` was created to hold child |
|
942 | // A node without a `DirstateEntry` was created to hold child | |
933 | // nodes, and is therefore a directory. |
|
943 | // nodes, and is therefore a directory. | |
934 | let state = node.state()?; |
|
944 | let state = node.state()?; | |
935 | Ok(state.is_none() && node.tracked_descendants_count() > 0) |
|
945 | Ok(state.is_none() && node.tracked_descendants_count() > 0) | |
936 | } else { |
|
946 | } else { | |
937 | Ok(false) |
|
947 | Ok(false) | |
938 | } |
|
948 | } | |
939 | }) |
|
949 | }) | |
940 | } |
|
950 | } | |
941 |
|
951 | |||
942 | pub fn has_dir( |
|
952 | pub fn has_dir( | |
943 | &mut self, |
|
953 | &mut self, | |
944 | directory: &HgPath, |
|
954 | directory: &HgPath, | |
945 | ) -> Result<bool, DirstateError> { |
|
955 | ) -> Result<bool, DirstateError> { | |
946 | self.with_dmap_mut(|map| { |
|
956 | self.with_dmap_mut(|map| { | |
947 | if let Some(node) = map.get_node(directory)? { |
|
957 | if let Some(node) = map.get_node(directory)? { | |
948 | // A node without a `DirstateEntry` was created to hold child |
|
958 | // A node without a `DirstateEntry` was created to hold child | |
949 | // nodes, and is therefore a directory. |
|
959 | // nodes, and is therefore a directory. | |
950 | let state = node.state()?; |
|
960 | let state = node.state()?; | |
951 | Ok(state.is_none() && node.descendants_with_entry_count() > 0) |
|
961 | Ok(state.is_none() && node.descendants_with_entry_count() > 0) | |
952 | } else { |
|
962 | } else { | |
953 | Ok(false) |
|
963 | Ok(false) | |
954 | } |
|
964 | } | |
955 | }) |
|
965 | }) | |
956 | } |
|
966 | } | |
957 |
|
967 | |||
958 | #[timed] |
|
968 | #[timed] | |
959 | pub fn pack_v1( |
|
969 | pub fn pack_v1( | |
960 | &self, |
|
970 | &self, | |
961 | parents: DirstateParents, |
|
971 | parents: DirstateParents, | |
962 | ) -> Result<Vec<u8>, DirstateError> { |
|
972 | ) -> Result<Vec<u8>, DirstateError> { | |
963 | let map = self.get_map(); |
|
973 | let map = self.get_map(); | |
964 | // Optizimation (to be measured?): pre-compute size to avoid `Vec` |
|
974 | // Optizimation (to be measured?): pre-compute size to avoid `Vec` | |
965 | // reallocations |
|
975 | // reallocations | |
966 | let mut size = parents.as_bytes().len(); |
|
976 | let mut size = parents.as_bytes().len(); | |
967 | for node in map.iter_nodes() { |
|
977 | for node in map.iter_nodes() { | |
968 | let node = node?; |
|
978 | let node = node?; | |
969 | if node.entry()?.is_some() { |
|
979 | if node.entry()?.is_some() { | |
970 | size += packed_entry_size( |
|
980 | size += packed_entry_size( | |
971 | node.full_path(map.on_disk)?, |
|
981 | node.full_path(map.on_disk)?, | |
972 | node.copy_source(map.on_disk)?, |
|
982 | node.copy_source(map.on_disk)?, | |
973 | ); |
|
983 | ); | |
974 | } |
|
984 | } | |
975 | } |
|
985 | } | |
976 |
|
986 | |||
977 | let mut packed = Vec::with_capacity(size); |
|
987 | let mut packed = Vec::with_capacity(size); | |
978 | packed.extend(parents.as_bytes()); |
|
988 | packed.extend(parents.as_bytes()); | |
979 |
|
989 | |||
980 | for node in map.iter_nodes() { |
|
990 | for node in map.iter_nodes() { | |
981 | let node = node?; |
|
991 | let node = node?; | |
982 | if let Some(entry) = node.entry()? { |
|
992 | if let Some(entry) = node.entry()? { | |
983 | pack_entry( |
|
993 | pack_entry( | |
984 | node.full_path(map.on_disk)?, |
|
994 | node.full_path(map.on_disk)?, | |
985 | &entry, |
|
995 | &entry, | |
986 | node.copy_source(map.on_disk)?, |
|
996 | node.copy_source(map.on_disk)?, | |
987 | &mut packed, |
|
997 | &mut packed, | |
988 | ); |
|
998 | ); | |
989 | } |
|
999 | } | |
990 | } |
|
1000 | } | |
991 | Ok(packed) |
|
1001 | Ok(packed) | |
992 | } |
|
1002 | } | |
993 |
|
1003 | |||
994 | /// Returns new data and metadata together with whether that data should be |
|
1004 | /// Returns new data and metadata together with whether that data should be | |
995 | /// appended to the existing data file whose content is at |
|
1005 | /// appended to the existing data file whose content is at | |
996 | /// `map.on_disk` (true), instead of written to a new data file |
|
1006 | /// `map.on_disk` (true), instead of written to a new data file | |
997 | /// (false), and the previous size of data on disk. |
|
1007 | /// (false), and the previous size of data on disk. | |
998 | #[timed] |
|
1008 | #[timed] | |
999 | pub fn pack_v2( |
|
1009 | pub fn pack_v2( | |
1000 | &self, |
|
1010 | &self, | |
1001 | can_append: bool, |
|
1011 | can_append: bool, | |
1002 | ) -> Result<(Vec<u8>, on_disk::TreeMetadata, bool, usize), DirstateError> |
|
1012 | ) -> Result<(Vec<u8>, on_disk::TreeMetadata, bool, usize), DirstateError> | |
1003 | { |
|
1013 | { | |
1004 | let map = self.get_map(); |
|
1014 | let map = self.get_map(); | |
1005 | on_disk::write(map, can_append) |
|
1015 | on_disk::write(map, can_append) | |
1006 | } |
|
1016 | } | |
1007 |
|
1017 | |||
1008 | /// `callback` allows the caller to process and do something with the |
|
1018 | /// `callback` allows the caller to process and do something with the | |
1009 | /// results of the status. This is needed to do so efficiently (i.e. |
|
1019 | /// results of the status. This is needed to do so efficiently (i.e. | |
1010 | /// without cloning the `DirstateStatus` object with its paths) because |
|
1020 | /// without cloning the `DirstateStatus` object with its paths) because | |
1011 | /// we need to borrow from `Self`. |
|
1021 | /// we need to borrow from `Self`. | |
1012 | pub fn with_status<R>( |
|
1022 | pub fn with_status<R>( | |
1013 | &mut self, |
|
1023 | &mut self, | |
1014 | matcher: &(dyn Matcher + Sync), |
|
1024 | matcher: &(dyn Matcher + Sync), | |
1015 | root_dir: PathBuf, |
|
1025 | root_dir: PathBuf, | |
1016 | ignore_files: Vec<PathBuf>, |
|
1026 | ignore_files: Vec<PathBuf>, | |
1017 | options: StatusOptions, |
|
1027 | options: StatusOptions, | |
1018 | callback: impl for<'r> FnOnce( |
|
1028 | callback: impl for<'r> FnOnce( | |
1019 | Result<(DirstateStatus<'r>, Vec<PatternFileWarning>), StatusError>, |
|
1029 | Result<(DirstateStatus<'r>, Vec<PatternFileWarning>), StatusError>, | |
1020 | ) -> R, |
|
1030 | ) -> R, | |
1021 | ) -> R { |
|
1031 | ) -> R { | |
1022 | self.with_dmap_mut(|map| { |
|
1032 | self.with_dmap_mut(|map| { | |
1023 | callback(super::status::status( |
|
1033 | callback(super::status::status( | |
1024 | map, |
|
1034 | map, | |
1025 | matcher, |
|
1035 | matcher, | |
1026 | root_dir, |
|
1036 | root_dir, | |
1027 | ignore_files, |
|
1037 | ignore_files, | |
1028 | options, |
|
1038 | options, | |
1029 | )) |
|
1039 | )) | |
1030 | }) |
|
1040 | }) | |
1031 | } |
|
1041 | } | |
1032 |
|
1042 | |||
1033 | pub fn copy_map_len(&self) -> usize { |
|
1043 | pub fn copy_map_len(&self) -> usize { | |
1034 | let map = self.get_map(); |
|
1044 | let map = self.get_map(); | |
1035 | map.nodes_with_copy_source_count as usize |
|
1045 | map.nodes_with_copy_source_count as usize | |
1036 | } |
|
1046 | } | |
1037 |
|
1047 | |||
1038 | pub fn copy_map_iter(&self) -> CopyMapIter<'_> { |
|
1048 | pub fn copy_map_iter(&self) -> CopyMapIter<'_> { | |
1039 | let map = self.get_map(); |
|
1049 | let map = self.get_map(); | |
1040 | Box::new(filter_map_results(map.iter_nodes(), move |node| { |
|
1050 | Box::new(filter_map_results(map.iter_nodes(), move |node| { | |
1041 | Ok(if let Some(source) = node.copy_source(map.on_disk)? { |
|
1051 | Ok(if let Some(source) = node.copy_source(map.on_disk)? { | |
1042 | Some((node.full_path(map.on_disk)?, source)) |
|
1052 | Some((node.full_path(map.on_disk)?, source)) | |
1043 | } else { |
|
1053 | } else { | |
1044 | None |
|
1054 | None | |
1045 | }) |
|
1055 | }) | |
1046 | })) |
|
1056 | })) | |
1047 | } |
|
1057 | } | |
1048 |
|
1058 | |||
1049 | pub fn copy_map_contains_key( |
|
1059 | pub fn copy_map_contains_key( | |
1050 | &self, |
|
1060 | &self, | |
1051 | key: &HgPath, |
|
1061 | key: &HgPath, | |
1052 | ) -> Result<bool, DirstateV2ParseError> { |
|
1062 | ) -> Result<bool, DirstateV2ParseError> { | |
1053 | let map = self.get_map(); |
|
1063 | let map = self.get_map(); | |
1054 | Ok(if let Some(node) = map.get_node(key)? { |
|
1064 | Ok(if let Some(node) = map.get_node(key)? { | |
1055 | node.has_copy_source() |
|
1065 | node.has_copy_source() | |
1056 | } else { |
|
1066 | } else { | |
1057 | false |
|
1067 | false | |
1058 | }) |
|
1068 | }) | |
1059 | } |
|
1069 | } | |
1060 |
|
1070 | |||
1061 | pub fn copy_map_get( |
|
1071 | pub fn copy_map_get( | |
1062 | &self, |
|
1072 | &self, | |
1063 | key: &HgPath, |
|
1073 | key: &HgPath, | |
1064 | ) -> Result<Option<&HgPath>, DirstateV2ParseError> { |
|
1074 | ) -> Result<Option<&HgPath>, DirstateV2ParseError> { | |
1065 | let map = self.get_map(); |
|
1075 | let map = self.get_map(); | |
1066 | if let Some(node) = map.get_node(key)? { |
|
1076 | if let Some(node) = map.get_node(key)? { | |
1067 | if let Some(source) = node.copy_source(map.on_disk)? { |
|
1077 | if let Some(source) = node.copy_source(map.on_disk)? { | |
1068 | return Ok(Some(source)); |
|
1078 | return Ok(Some(source)); | |
1069 | } |
|
1079 | } | |
1070 | } |
|
1080 | } | |
1071 | Ok(None) |
|
1081 | Ok(None) | |
1072 | } |
|
1082 | } | |
1073 |
|
1083 | |||
1074 | pub fn copy_map_remove( |
|
1084 | pub fn copy_map_remove( | |
1075 | &mut self, |
|
1085 | &mut self, | |
1076 | key: &HgPath, |
|
1086 | key: &HgPath, | |
1077 | ) -> Result<Option<HgPathBuf>, DirstateV2ParseError> { |
|
1087 | ) -> Result<Option<HgPathBuf>, DirstateV2ParseError> { | |
1078 | self.with_dmap_mut(|map| { |
|
1088 | self.with_dmap_mut(|map| { | |
1079 | let count = &mut map.nodes_with_copy_source_count; |
|
1089 | let count = &mut map.nodes_with_copy_source_count; | |
1080 | let unreachable_bytes = &mut map.unreachable_bytes; |
|
1090 | let unreachable_bytes = &mut map.unreachable_bytes; | |
1081 | Ok(DirstateMap::get_node_mut( |
|
1091 | Ok(DirstateMap::get_node_mut( | |
1082 | map.on_disk, |
|
1092 | map.on_disk, | |
1083 | unreachable_bytes, |
|
1093 | unreachable_bytes, | |
1084 | &mut map.root, |
|
1094 | &mut map.root, | |
1085 | key, |
|
1095 | key, | |
1086 | )? |
|
1096 | )? | |
1087 | .and_then(|node| { |
|
1097 | .and_then(|node| { | |
1088 | if let Some(source) = &node.copy_source { |
|
1098 | if let Some(source) = &node.copy_source { | |
1089 | *count -= 1; |
|
1099 | *count -= 1; | |
1090 | DirstateMap::count_dropped_path(unreachable_bytes, source); |
|
1100 | DirstateMap::count_dropped_path(unreachable_bytes, source); | |
1091 | } |
|
1101 | } | |
1092 | node.copy_source.take().map(Cow::into_owned) |
|
1102 | node.copy_source.take().map(Cow::into_owned) | |
1093 | })) |
|
1103 | })) | |
1094 | }) |
|
1104 | }) | |
1095 | } |
|
1105 | } | |
1096 |
|
1106 | |||
1097 | pub fn copy_map_insert( |
|
1107 | pub fn copy_map_insert( | |
1098 | &mut self, |
|
1108 | &mut self, | |
1099 | key: HgPathBuf, |
|
1109 | key: HgPathBuf, | |
1100 | value: HgPathBuf, |
|
1110 | value: HgPathBuf, | |
1101 | ) -> Result<Option<HgPathBuf>, DirstateV2ParseError> { |
|
1111 | ) -> Result<Option<HgPathBuf>, DirstateV2ParseError> { | |
1102 | self.with_dmap_mut(|map| { |
|
1112 | self.with_dmap_mut(|map| { | |
1103 | let node = DirstateMap::get_or_insert_node( |
|
1113 | let node = DirstateMap::get_or_insert_node( | |
1104 | map.on_disk, |
|
1114 | map.on_disk, | |
1105 | &mut map.unreachable_bytes, |
|
1115 | &mut map.unreachable_bytes, | |
1106 | &mut map.root, |
|
1116 | &mut map.root, | |
1107 | &key, |
|
1117 | &key, | |
1108 | WithBasename::to_cow_owned, |
|
1118 | WithBasename::to_cow_owned, | |
1109 | |_ancestor| {}, |
|
1119 | |_ancestor| {}, | |
1110 | )?; |
|
1120 | )?; | |
1111 | if node.copy_source.is_none() { |
|
1121 | if node.copy_source.is_none() { | |
1112 | map.nodes_with_copy_source_count += 1 |
|
1122 | map.nodes_with_copy_source_count += 1 | |
1113 | } |
|
1123 | } | |
1114 | Ok(node.copy_source.replace(value.into()).map(Cow::into_owned)) |
|
1124 | Ok(node.copy_source.replace(value.into()).map(Cow::into_owned)) | |
1115 | }) |
|
1125 | }) | |
1116 | } |
|
1126 | } | |
1117 |
|
1127 | |||
1118 | pub fn len(&self) -> usize { |
|
1128 | pub fn len(&self) -> usize { | |
1119 | let map = self.get_map(); |
|
1129 | let map = self.get_map(); | |
1120 | map.nodes_with_entry_count as usize |
|
1130 | map.nodes_with_entry_count as usize | |
1121 | } |
|
1131 | } | |
1122 |
|
1132 | |||
1123 | pub fn contains_key( |
|
1133 | pub fn contains_key( | |
1124 | &self, |
|
1134 | &self, | |
1125 | key: &HgPath, |
|
1135 | key: &HgPath, | |
1126 | ) -> Result<bool, DirstateV2ParseError> { |
|
1136 | ) -> Result<bool, DirstateV2ParseError> { | |
1127 | Ok(self.get(key)?.is_some()) |
|
1137 | Ok(self.get(key)?.is_some()) | |
1128 | } |
|
1138 | } | |
1129 |
|
1139 | |||
1130 | pub fn get( |
|
1140 | pub fn get( | |
1131 | &self, |
|
1141 | &self, | |
1132 | key: &HgPath, |
|
1142 | key: &HgPath, | |
1133 | ) -> Result<Option<DirstateEntry>, DirstateV2ParseError> { |
|
1143 | ) -> Result<Option<DirstateEntry>, DirstateV2ParseError> { | |
1134 | let map = self.get_map(); |
|
1144 | let map = self.get_map(); | |
1135 | Ok(if let Some(node) = map.get_node(key)? { |
|
1145 | Ok(if let Some(node) = map.get_node(key)? { | |
1136 | node.entry()? |
|
1146 | node.entry()? | |
1137 | } else { |
|
1147 | } else { | |
1138 | None |
|
1148 | None | |
1139 | }) |
|
1149 | }) | |
1140 | } |
|
1150 | } | |
1141 |
|
1151 | |||
1142 | pub fn iter(&self) -> StateMapIter<'_> { |
|
1152 | pub fn iter(&self) -> StateMapIter<'_> { | |
1143 | let map = self.get_map(); |
|
1153 | let map = self.get_map(); | |
1144 | Box::new(filter_map_results(map.iter_nodes(), move |node| { |
|
1154 | Box::new(filter_map_results(map.iter_nodes(), move |node| { | |
1145 | Ok(if let Some(entry) = node.entry()? { |
|
1155 | Ok(if let Some(entry) = node.entry()? { | |
1146 | Some((node.full_path(map.on_disk)?, entry)) |
|
1156 | Some((node.full_path(map.on_disk)?, entry)) | |
1147 | } else { |
|
1157 | } else { | |
1148 | None |
|
1158 | None | |
1149 | }) |
|
1159 | }) | |
1150 | })) |
|
1160 | })) | |
1151 | } |
|
1161 | } | |
1152 |
|
1162 | |||
1153 | pub fn iter_tracked_dirs( |
|
1163 | pub fn iter_tracked_dirs( | |
1154 | &mut self, |
|
1164 | &mut self, | |
1155 | ) -> Result< |
|
1165 | ) -> Result< | |
1156 | Box< |
|
1166 | Box< | |
1157 | dyn Iterator<Item = Result<&HgPath, DirstateV2ParseError>> |
|
1167 | dyn Iterator<Item = Result<&HgPath, DirstateV2ParseError>> | |
1158 | + Send |
|
1168 | + Send | |
1159 | + '_, |
|
1169 | + '_, | |
1160 | >, |
|
1170 | >, | |
1161 | DirstateError, |
|
1171 | DirstateError, | |
1162 | > { |
|
1172 | > { | |
1163 | let map = self.get_map(); |
|
1173 | let map = self.get_map(); | |
1164 | let on_disk = map.on_disk; |
|
1174 | let on_disk = map.on_disk; | |
1165 | Ok(Box::new(filter_map_results( |
|
1175 | Ok(Box::new(filter_map_results( | |
1166 | map.iter_nodes(), |
|
1176 | map.iter_nodes(), | |
1167 | move |node| { |
|
1177 | move |node| { | |
1168 | Ok(if node.tracked_descendants_count() > 0 { |
|
1178 | Ok(if node.tracked_descendants_count() > 0 { | |
1169 | Some(node.full_path(on_disk)?) |
|
1179 | Some(node.full_path(on_disk)?) | |
1170 | } else { |
|
1180 | } else { | |
1171 | None |
|
1181 | None | |
1172 | }) |
|
1182 | }) | |
1173 | }, |
|
1183 | }, | |
1174 | ))) |
|
1184 | ))) | |
1175 | } |
|
1185 | } | |
1176 |
|
1186 | |||
1177 | pub fn debug_iter( |
|
1187 | pub fn debug_iter( | |
1178 | &self, |
|
1188 | &self, | |
1179 | all: bool, |
|
1189 | all: bool, | |
1180 | ) -> Box< |
|
1190 | ) -> Box< | |
1181 | dyn Iterator< |
|
1191 | dyn Iterator< | |
1182 | Item = Result< |
|
1192 | Item = Result< | |
1183 | (&HgPath, (u8, i32, i32, i32)), |
|
1193 | (&HgPath, (u8, i32, i32, i32)), | |
1184 | DirstateV2ParseError, |
|
1194 | DirstateV2ParseError, | |
1185 | >, |
|
1195 | >, | |
1186 | > + Send |
|
1196 | > + Send | |
1187 | + '_, |
|
1197 | + '_, | |
1188 | > { |
|
1198 | > { | |
1189 | let map = self.get_map(); |
|
1199 | let map = self.get_map(); | |
1190 | Box::new(filter_map_results(map.iter_nodes(), move |node| { |
|
1200 | Box::new(filter_map_results(map.iter_nodes(), move |node| { | |
1191 | let debug_tuple = if let Some(entry) = node.entry()? { |
|
1201 | let debug_tuple = if let Some(entry) = node.entry()? { | |
1192 | entry.debug_tuple() |
|
1202 | entry.debug_tuple() | |
1193 | } else if !all { |
|
1203 | } else if !all { | |
1194 | return Ok(None); |
|
1204 | return Ok(None); | |
1195 | } else if let Some(mtime) = node.cached_directory_mtime()? { |
|
1205 | } else if let Some(mtime) = node.cached_directory_mtime()? { | |
1196 | (b' ', 0, -1, mtime.truncated_seconds() as i32) |
|
1206 | (b' ', 0, -1, mtime.truncated_seconds() as i32) | |
1197 | } else { |
|
1207 | } else { | |
1198 | (b' ', 0, -1, -1) |
|
1208 | (b' ', 0, -1, -1) | |
1199 | }; |
|
1209 | }; | |
1200 | Ok(Some((node.full_path(map.on_disk)?, debug_tuple))) |
|
1210 | Ok(Some((node.full_path(map.on_disk)?, debug_tuple))) | |
1201 | })) |
|
1211 | })) | |
1202 | } |
|
1212 | } | |
1203 | } |
|
1213 | } |
@@ -1,849 +1,853 b'' | |||||
1 | //! The "version 2" disk representation of the dirstate |
|
1 | //! The "version 2" disk representation of the dirstate | |
2 | //! |
|
2 | //! | |
3 | //! See `mercurial/helptext/internals/dirstate-v2.txt` |
|
3 | //! See `mercurial/helptext/internals/dirstate-v2.txt` | |
4 |
|
4 | |||
5 | use crate::dirstate::TruncatedTimestamp; |
|
5 | use crate::dirstate::TruncatedTimestamp; | |
|
6 | use crate::dirstate_tree::dirstate_map::DirstateVersion; | |||
6 | use crate::dirstate_tree::dirstate_map::{self, DirstateMap, NodeRef}; |
|
7 | use crate::dirstate_tree::dirstate_map::{self, DirstateMap, NodeRef}; | |
7 | use crate::dirstate_tree::path_with_basename::WithBasename; |
|
8 | use crate::dirstate_tree::path_with_basename::WithBasename; | |
8 | use crate::errors::HgError; |
|
9 | use crate::errors::HgError; | |
9 | use crate::utils::hg_path::HgPath; |
|
10 | use crate::utils::hg_path::HgPath; | |
10 | use crate::DirstateEntry; |
|
11 | use crate::DirstateEntry; | |
11 | use crate::DirstateError; |
|
12 | use crate::DirstateError; | |
12 | use crate::DirstateParents; |
|
13 | use crate::DirstateParents; | |
13 | use bitflags::bitflags; |
|
14 | use bitflags::bitflags; | |
14 | use bytes_cast::unaligned::{U16Be, U32Be}; |
|
15 | use bytes_cast::unaligned::{U16Be, U32Be}; | |
15 | use bytes_cast::BytesCast; |
|
16 | use bytes_cast::BytesCast; | |
16 | use format_bytes::format_bytes; |
|
17 | use format_bytes::format_bytes; | |
17 | use rand::Rng; |
|
18 | use rand::Rng; | |
18 | use std::borrow::Cow; |
|
19 | use std::borrow::Cow; | |
19 | use std::convert::{TryFrom, TryInto}; |
|
20 | use std::convert::{TryFrom, TryInto}; | |
20 | use std::fmt::Write; |
|
21 | use std::fmt::Write; | |
21 |
|
22 | |||
22 | /// Added at the start of `.hg/dirstate` when the "v2" format is used. |
|
23 | /// Added at the start of `.hg/dirstate` when the "v2" format is used. | |
23 | /// This a redundant sanity check more than an actual "magic number" since |
|
24 | /// This a redundant sanity check more than an actual "magic number" since | |
24 | /// `.hg/requires` already governs which format should be used. |
|
25 | /// `.hg/requires` already governs which format should be used. | |
25 | pub const V2_FORMAT_MARKER: &[u8; 12] = b"dirstate-v2\n"; |
|
26 | pub const V2_FORMAT_MARKER: &[u8; 12] = b"dirstate-v2\n"; | |
26 |
|
27 | |||
27 | /// Keep space for 256-bit hashes |
|
28 | /// Keep space for 256-bit hashes | |
28 | const STORED_NODE_ID_BYTES: usize = 32; |
|
29 | const STORED_NODE_ID_BYTES: usize = 32; | |
29 |
|
30 | |||
30 | /// … even though only 160 bits are used for now, with SHA-1 |
|
31 | /// … even though only 160 bits are used for now, with SHA-1 | |
31 | const USED_NODE_ID_BYTES: usize = 20; |
|
32 | const USED_NODE_ID_BYTES: usize = 20; | |
32 |
|
33 | |||
33 | pub(super) const IGNORE_PATTERNS_HASH_LEN: usize = 20; |
|
34 | pub(super) const IGNORE_PATTERNS_HASH_LEN: usize = 20; | |
34 | pub(super) type IgnorePatternsHash = [u8; IGNORE_PATTERNS_HASH_LEN]; |
|
35 | pub(super) type IgnorePatternsHash = [u8; IGNORE_PATTERNS_HASH_LEN]; | |
35 |
|
36 | |||
36 | /// Must match constants of the same names in `mercurial/dirstateutils/v2.py` |
|
37 | /// Must match constants of the same names in `mercurial/dirstateutils/v2.py` | |
37 | const TREE_METADATA_SIZE: usize = 44; |
|
38 | const TREE_METADATA_SIZE: usize = 44; | |
38 | const NODE_SIZE: usize = 44; |
|
39 | const NODE_SIZE: usize = 44; | |
39 |
|
40 | |||
40 | /// Make sure that size-affecting changes are made knowingly |
|
41 | /// Make sure that size-affecting changes are made knowingly | |
41 | #[allow(unused)] |
|
42 | #[allow(unused)] | |
42 | fn static_assert_size_of() { |
|
43 | fn static_assert_size_of() { | |
43 | let _ = std::mem::transmute::<TreeMetadata, [u8; TREE_METADATA_SIZE]>; |
|
44 | let _ = std::mem::transmute::<TreeMetadata, [u8; TREE_METADATA_SIZE]>; | |
44 | let _ = std::mem::transmute::<DocketHeader, [u8; TREE_METADATA_SIZE + 81]>; |
|
45 | let _ = std::mem::transmute::<DocketHeader, [u8; TREE_METADATA_SIZE + 81]>; | |
45 | let _ = std::mem::transmute::<Node, [u8; NODE_SIZE]>; |
|
46 | let _ = std::mem::transmute::<Node, [u8; NODE_SIZE]>; | |
46 | } |
|
47 | } | |
47 |
|
48 | |||
48 | // Must match `HEADER` in `mercurial/dirstateutils/docket.py` |
|
49 | // Must match `HEADER` in `mercurial/dirstateutils/docket.py` | |
49 | #[derive(BytesCast)] |
|
50 | #[derive(BytesCast)] | |
50 | #[repr(C)] |
|
51 | #[repr(C)] | |
51 | struct DocketHeader { |
|
52 | struct DocketHeader { | |
52 | marker: [u8; V2_FORMAT_MARKER.len()], |
|
53 | marker: [u8; V2_FORMAT_MARKER.len()], | |
53 | parent_1: [u8; STORED_NODE_ID_BYTES], |
|
54 | parent_1: [u8; STORED_NODE_ID_BYTES], | |
54 | parent_2: [u8; STORED_NODE_ID_BYTES], |
|
55 | parent_2: [u8; STORED_NODE_ID_BYTES], | |
55 |
|
56 | |||
56 | metadata: TreeMetadata, |
|
57 | metadata: TreeMetadata, | |
57 |
|
58 | |||
58 | /// Counted in bytes |
|
59 | /// Counted in bytes | |
59 | data_size: Size, |
|
60 | data_size: Size, | |
60 |
|
61 | |||
61 | uuid_size: u8, |
|
62 | uuid_size: u8, | |
62 | } |
|
63 | } | |
63 |
|
64 | |||
64 | pub struct Docket<'on_disk> { |
|
65 | pub struct Docket<'on_disk> { | |
65 | header: &'on_disk DocketHeader, |
|
66 | header: &'on_disk DocketHeader, | |
66 | pub uuid: &'on_disk [u8], |
|
67 | pub uuid: &'on_disk [u8], | |
67 | } |
|
68 | } | |
68 |
|
69 | |||
69 | /// Fields are documented in the *Tree metadata in the docket file* |
|
70 | /// Fields are documented in the *Tree metadata in the docket file* | |
70 | /// section of `mercurial/helptext/internals/dirstate-v2.txt` |
|
71 | /// section of `mercurial/helptext/internals/dirstate-v2.txt` | |
71 | #[derive(BytesCast)] |
|
72 | #[derive(BytesCast)] | |
72 | #[repr(C)] |
|
73 | #[repr(C)] | |
73 | pub struct TreeMetadata { |
|
74 | pub struct TreeMetadata { | |
74 | root_nodes: ChildNodes, |
|
75 | root_nodes: ChildNodes, | |
75 | nodes_with_entry_count: Size, |
|
76 | nodes_with_entry_count: Size, | |
76 | nodes_with_copy_source_count: Size, |
|
77 | nodes_with_copy_source_count: Size, | |
77 | unreachable_bytes: Size, |
|
78 | unreachable_bytes: Size, | |
78 | unused: [u8; 4], |
|
79 | unused: [u8; 4], | |
79 |
|
80 | |||
80 | /// See *Optional hash of ignore patterns* section of |
|
81 | /// See *Optional hash of ignore patterns* section of | |
81 | /// `mercurial/helptext/internals/dirstate-v2.txt` |
|
82 | /// `mercurial/helptext/internals/dirstate-v2.txt` | |
82 | ignore_patterns_hash: IgnorePatternsHash, |
|
83 | ignore_patterns_hash: IgnorePatternsHash, | |
83 | } |
|
84 | } | |
84 |
|
85 | |||
85 | /// Fields are documented in the *The data file format* |
|
86 | /// Fields are documented in the *The data file format* | |
86 | /// section of `mercurial/helptext/internals/dirstate-v2.txt` |
|
87 | /// section of `mercurial/helptext/internals/dirstate-v2.txt` | |
87 | #[derive(BytesCast)] |
|
88 | #[derive(BytesCast)] | |
88 | #[repr(C)] |
|
89 | #[repr(C)] | |
89 | pub(super) struct Node { |
|
90 | pub(super) struct Node { | |
90 | full_path: PathSlice, |
|
91 | full_path: PathSlice, | |
91 |
|
92 | |||
92 | /// In bytes from `self.full_path.start` |
|
93 | /// In bytes from `self.full_path.start` | |
93 | base_name_start: PathSize, |
|
94 | base_name_start: PathSize, | |
94 |
|
95 | |||
95 | copy_source: OptPathSlice, |
|
96 | copy_source: OptPathSlice, | |
96 | children: ChildNodes, |
|
97 | children: ChildNodes, | |
97 | pub(super) descendants_with_entry_count: Size, |
|
98 | pub(super) descendants_with_entry_count: Size, | |
98 | pub(super) tracked_descendants_count: Size, |
|
99 | pub(super) tracked_descendants_count: Size, | |
99 | flags: U16Be, |
|
100 | flags: U16Be, | |
100 | size: U32Be, |
|
101 | size: U32Be, | |
101 | mtime: PackedTruncatedTimestamp, |
|
102 | mtime: PackedTruncatedTimestamp, | |
102 | } |
|
103 | } | |
103 |
|
104 | |||
104 | bitflags! { |
|
105 | bitflags! { | |
105 | #[repr(C)] |
|
106 | #[repr(C)] | |
106 | struct Flags: u16 { |
|
107 | struct Flags: u16 { | |
107 | const WDIR_TRACKED = 1 << 0; |
|
108 | const WDIR_TRACKED = 1 << 0; | |
108 | const P1_TRACKED = 1 << 1; |
|
109 | const P1_TRACKED = 1 << 1; | |
109 | const P2_INFO = 1 << 2; |
|
110 | const P2_INFO = 1 << 2; | |
110 | const MODE_EXEC_PERM = 1 << 3; |
|
111 | const MODE_EXEC_PERM = 1 << 3; | |
111 | const MODE_IS_SYMLINK = 1 << 4; |
|
112 | const MODE_IS_SYMLINK = 1 << 4; | |
112 | const HAS_FALLBACK_EXEC = 1 << 5; |
|
113 | const HAS_FALLBACK_EXEC = 1 << 5; | |
113 | const FALLBACK_EXEC = 1 << 6; |
|
114 | const FALLBACK_EXEC = 1 << 6; | |
114 | const HAS_FALLBACK_SYMLINK = 1 << 7; |
|
115 | const HAS_FALLBACK_SYMLINK = 1 << 7; | |
115 | const FALLBACK_SYMLINK = 1 << 8; |
|
116 | const FALLBACK_SYMLINK = 1 << 8; | |
116 | const EXPECTED_STATE_IS_MODIFIED = 1 << 9; |
|
117 | const EXPECTED_STATE_IS_MODIFIED = 1 << 9; | |
117 | const HAS_MODE_AND_SIZE = 1 <<10; |
|
118 | const HAS_MODE_AND_SIZE = 1 <<10; | |
118 | const HAS_MTIME = 1 <<11; |
|
119 | const HAS_MTIME = 1 <<11; | |
119 | const MTIME_SECOND_AMBIGUOUS = 1 << 12; |
|
120 | const MTIME_SECOND_AMBIGUOUS = 1 << 12; | |
120 | const DIRECTORY = 1 <<13; |
|
121 | const DIRECTORY = 1 <<13; | |
121 | const ALL_UNKNOWN_RECORDED = 1 <<14; |
|
122 | const ALL_UNKNOWN_RECORDED = 1 <<14; | |
122 | const ALL_IGNORED_RECORDED = 1 <<15; |
|
123 | const ALL_IGNORED_RECORDED = 1 <<15; | |
123 | } |
|
124 | } | |
124 | } |
|
125 | } | |
125 |
|
126 | |||
126 | /// Duration since the Unix epoch |
|
127 | /// Duration since the Unix epoch | |
127 | #[derive(BytesCast, Copy, Clone)] |
|
128 | #[derive(BytesCast, Copy, Clone)] | |
128 | #[repr(C)] |
|
129 | #[repr(C)] | |
129 | struct PackedTruncatedTimestamp { |
|
130 | struct PackedTruncatedTimestamp { | |
130 | truncated_seconds: U32Be, |
|
131 | truncated_seconds: U32Be, | |
131 | nanoseconds: U32Be, |
|
132 | nanoseconds: U32Be, | |
132 | } |
|
133 | } | |
133 |
|
134 | |||
134 | /// Counted in bytes from the start of the file |
|
135 | /// Counted in bytes from the start of the file | |
135 | /// |
|
136 | /// | |
136 | /// NOTE: not supporting `.hg/dirstate` files larger than 4 GiB. |
|
137 | /// NOTE: not supporting `.hg/dirstate` files larger than 4 GiB. | |
137 | type Offset = U32Be; |
|
138 | type Offset = U32Be; | |
138 |
|
139 | |||
139 | /// Counted in number of items |
|
140 | /// Counted in number of items | |
140 | /// |
|
141 | /// | |
141 | /// NOTE: we choose not to support counting more than 4 billion nodes anywhere. |
|
142 | /// NOTE: we choose not to support counting more than 4 billion nodes anywhere. | |
142 | type Size = U32Be; |
|
143 | type Size = U32Be; | |
143 |
|
144 | |||
144 | /// Counted in bytes |
|
145 | /// Counted in bytes | |
145 | /// |
|
146 | /// | |
146 | /// NOTE: we choose not to support file names/paths longer than 64 KiB. |
|
147 | /// NOTE: we choose not to support file names/paths longer than 64 KiB. | |
147 | type PathSize = U16Be; |
|
148 | type PathSize = U16Be; | |
148 |
|
149 | |||
149 | /// A contiguous sequence of `len` times `Node`, representing the child nodes |
|
150 | /// A contiguous sequence of `len` times `Node`, representing the child nodes | |
150 | /// of either some other node or of the repository root. |
|
151 | /// of either some other node or of the repository root. | |
151 | /// |
|
152 | /// | |
152 | /// Always sorted by ascending `full_path`, to allow binary search. |
|
153 | /// Always sorted by ascending `full_path`, to allow binary search. | |
153 | /// Since nodes with the same parent nodes also have the same parent path, |
|
154 | /// Since nodes with the same parent nodes also have the same parent path, | |
154 | /// only the `base_name`s need to be compared during binary search. |
|
155 | /// only the `base_name`s need to be compared during binary search. | |
155 | #[derive(BytesCast, Copy, Clone)] |
|
156 | #[derive(BytesCast, Copy, Clone)] | |
156 | #[repr(C)] |
|
157 | #[repr(C)] | |
157 | struct ChildNodes { |
|
158 | struct ChildNodes { | |
158 | start: Offset, |
|
159 | start: Offset, | |
159 | len: Size, |
|
160 | len: Size, | |
160 | } |
|
161 | } | |
161 |
|
162 | |||
162 | /// A `HgPath` of `len` bytes |
|
163 | /// A `HgPath` of `len` bytes | |
163 | #[derive(BytesCast, Copy, Clone)] |
|
164 | #[derive(BytesCast, Copy, Clone)] | |
164 | #[repr(C)] |
|
165 | #[repr(C)] | |
165 | struct PathSlice { |
|
166 | struct PathSlice { | |
166 | start: Offset, |
|
167 | start: Offset, | |
167 | len: PathSize, |
|
168 | len: PathSize, | |
168 | } |
|
169 | } | |
169 |
|
170 | |||
170 | /// Either nothing if `start == 0`, or a `HgPath` of `len` bytes |
|
171 | /// Either nothing if `start == 0`, or a `HgPath` of `len` bytes | |
171 | type OptPathSlice = PathSlice; |
|
172 | type OptPathSlice = PathSlice; | |
172 |
|
173 | |||
173 | /// Unexpected file format found in `.hg/dirstate` with the "v2" format. |
|
174 | /// Unexpected file format found in `.hg/dirstate` with the "v2" format. | |
174 | /// |
|
175 | /// | |
175 | /// This should only happen if Mercurial is buggy or a repository is corrupted. |
|
176 | /// This should only happen if Mercurial is buggy or a repository is corrupted. | |
176 | #[derive(Debug)] |
|
177 | #[derive(Debug)] | |
177 | pub struct DirstateV2ParseError; |
|
178 | pub struct DirstateV2ParseError; | |
178 |
|
179 | |||
179 | impl From<DirstateV2ParseError> for HgError { |
|
180 | impl From<DirstateV2ParseError> for HgError { | |
180 | fn from(_: DirstateV2ParseError) -> Self { |
|
181 | fn from(_: DirstateV2ParseError) -> Self { | |
181 | HgError::corrupted("dirstate-v2 parse error") |
|
182 | HgError::corrupted("dirstate-v2 parse error") | |
182 | } |
|
183 | } | |
183 | } |
|
184 | } | |
184 |
|
185 | |||
185 | impl From<DirstateV2ParseError> for crate::DirstateError { |
|
186 | impl From<DirstateV2ParseError> for crate::DirstateError { | |
186 | fn from(error: DirstateV2ParseError) -> Self { |
|
187 | fn from(error: DirstateV2ParseError) -> Self { | |
187 | HgError::from(error).into() |
|
188 | HgError::from(error).into() | |
188 | } |
|
189 | } | |
189 | } |
|
190 | } | |
190 |
|
191 | |||
191 | impl TreeMetadata { |
|
192 | impl TreeMetadata { | |
192 | pub fn as_bytes(&self) -> &[u8] { |
|
193 | pub fn as_bytes(&self) -> &[u8] { | |
193 | BytesCast::as_bytes(self) |
|
194 | BytesCast::as_bytes(self) | |
194 | } |
|
195 | } | |
195 | } |
|
196 | } | |
196 |
|
197 | |||
197 | impl<'on_disk> Docket<'on_disk> { |
|
198 | impl<'on_disk> Docket<'on_disk> { | |
198 | /// Generate the identifier for a new data file |
|
199 | /// Generate the identifier for a new data file | |
199 | /// |
|
200 | /// | |
200 | /// TODO: support the `HGTEST_UUIDFILE` environment variable. |
|
201 | /// TODO: support the `HGTEST_UUIDFILE` environment variable. | |
201 | /// See `mercurial/revlogutils/docket.py` |
|
202 | /// See `mercurial/revlogutils/docket.py` | |
202 | pub fn new_uid() -> String { |
|
203 | pub fn new_uid() -> String { | |
203 | const ID_LENGTH: usize = 8; |
|
204 | const ID_LENGTH: usize = 8; | |
204 | let mut id = String::with_capacity(ID_LENGTH); |
|
205 | let mut id = String::with_capacity(ID_LENGTH); | |
205 | let mut rng = rand::thread_rng(); |
|
206 | let mut rng = rand::thread_rng(); | |
206 | for _ in 0..ID_LENGTH { |
|
207 | for _ in 0..ID_LENGTH { | |
207 | // One random hexadecimal digit. |
|
208 | // One random hexadecimal digit. | |
208 | // `unwrap` never panics because `impl Write for String` |
|
209 | // `unwrap` never panics because `impl Write for String` | |
209 | // never returns an error. |
|
210 | // never returns an error. | |
210 | write!(&mut id, "{:x}", rng.gen_range(0..16)).unwrap(); |
|
211 | write!(&mut id, "{:x}", rng.gen_range(0..16)).unwrap(); | |
211 | } |
|
212 | } | |
212 | id |
|
213 | id | |
213 | } |
|
214 | } | |
214 |
|
215 | |||
215 | pub fn serialize( |
|
216 | pub fn serialize( | |
216 | parents: DirstateParents, |
|
217 | parents: DirstateParents, | |
217 | tree_metadata: TreeMetadata, |
|
218 | tree_metadata: TreeMetadata, | |
218 | data_size: u64, |
|
219 | data_size: u64, | |
219 | uuid: &[u8], |
|
220 | uuid: &[u8], | |
220 | ) -> Result<Vec<u8>, std::num::TryFromIntError> { |
|
221 | ) -> Result<Vec<u8>, std::num::TryFromIntError> { | |
221 | let header = DocketHeader { |
|
222 | let header = DocketHeader { | |
222 | marker: *V2_FORMAT_MARKER, |
|
223 | marker: *V2_FORMAT_MARKER, | |
223 | parent_1: parents.p1.pad_to_256_bits(), |
|
224 | parent_1: parents.p1.pad_to_256_bits(), | |
224 | parent_2: parents.p2.pad_to_256_bits(), |
|
225 | parent_2: parents.p2.pad_to_256_bits(), | |
225 | metadata: tree_metadata, |
|
226 | metadata: tree_metadata, | |
226 | data_size: u32::try_from(data_size)?.into(), |
|
227 | data_size: u32::try_from(data_size)?.into(), | |
227 | uuid_size: uuid.len().try_into()?, |
|
228 | uuid_size: uuid.len().try_into()?, | |
228 | }; |
|
229 | }; | |
229 | let header = header.as_bytes(); |
|
230 | let header = header.as_bytes(); | |
230 | let mut docket = Vec::with_capacity(header.len() + uuid.len()); |
|
231 | let mut docket = Vec::with_capacity(header.len() + uuid.len()); | |
231 | docket.extend_from_slice(header); |
|
232 | docket.extend_from_slice(header); | |
232 | docket.extend_from_slice(uuid); |
|
233 | docket.extend_from_slice(uuid); | |
233 | Ok(docket) |
|
234 | Ok(docket) | |
234 | } |
|
235 | } | |
235 |
|
236 | |||
236 | pub fn parents(&self) -> DirstateParents { |
|
237 | pub fn parents(&self) -> DirstateParents { | |
237 | use crate::Node; |
|
238 | use crate::Node; | |
238 | let p1 = Node::try_from(&self.header.parent_1[..USED_NODE_ID_BYTES]) |
|
239 | let p1 = Node::try_from(&self.header.parent_1[..USED_NODE_ID_BYTES]) | |
239 | .unwrap() |
|
240 | .unwrap() | |
240 | .clone(); |
|
241 | .clone(); | |
241 | let p2 = Node::try_from(&self.header.parent_2[..USED_NODE_ID_BYTES]) |
|
242 | let p2 = Node::try_from(&self.header.parent_2[..USED_NODE_ID_BYTES]) | |
242 | .unwrap() |
|
243 | .unwrap() | |
243 | .clone(); |
|
244 | .clone(); | |
244 | DirstateParents { p1, p2 } |
|
245 | DirstateParents { p1, p2 } | |
245 | } |
|
246 | } | |
246 |
|
247 | |||
247 | pub fn tree_metadata(&self) -> &[u8] { |
|
248 | pub fn tree_metadata(&self) -> &[u8] { | |
248 | self.header.metadata.as_bytes() |
|
249 | self.header.metadata.as_bytes() | |
249 | } |
|
250 | } | |
250 |
|
251 | |||
251 | pub fn data_size(&self) -> usize { |
|
252 | pub fn data_size(&self) -> usize { | |
252 | // This `unwrap` could only panic on a 16-bit CPU |
|
253 | // This `unwrap` could only panic on a 16-bit CPU | |
253 | self.header.data_size.get().try_into().unwrap() |
|
254 | self.header.data_size.get().try_into().unwrap() | |
254 | } |
|
255 | } | |
255 |
|
256 | |||
256 | pub fn data_filename(&self) -> String { |
|
257 | pub fn data_filename(&self) -> String { | |
257 | String::from_utf8(format_bytes!(b"dirstate.{}", self.uuid)).unwrap() |
|
258 | String::from_utf8(format_bytes!(b"dirstate.{}", self.uuid)).unwrap() | |
258 | } |
|
259 | } | |
259 | } |
|
260 | } | |
260 |
|
261 | |||
261 | pub fn read_docket( |
|
262 | pub fn read_docket( | |
262 | on_disk: &[u8], |
|
263 | on_disk: &[u8], | |
263 | ) -> Result<Docket<'_>, DirstateV2ParseError> { |
|
264 | ) -> Result<Docket<'_>, DirstateV2ParseError> { | |
264 | let (header, uuid) = |
|
265 | let (header, uuid) = | |
265 | DocketHeader::from_bytes(on_disk).map_err(|_| DirstateV2ParseError)?; |
|
266 | DocketHeader::from_bytes(on_disk).map_err(|_| DirstateV2ParseError)?; | |
266 | let uuid_size = header.uuid_size as usize; |
|
267 | let uuid_size = header.uuid_size as usize; | |
267 | if header.marker == *V2_FORMAT_MARKER && uuid.len() == uuid_size { |
|
268 | if header.marker == *V2_FORMAT_MARKER && uuid.len() == uuid_size { | |
268 | Ok(Docket { header, uuid }) |
|
269 | Ok(Docket { header, uuid }) | |
269 | } else { |
|
270 | } else { | |
270 | Err(DirstateV2ParseError) |
|
271 | Err(DirstateV2ParseError) | |
271 | } |
|
272 | } | |
272 | } |
|
273 | } | |
273 |
|
274 | |||
274 | pub(super) fn read<'on_disk>( |
|
275 | pub(super) fn read<'on_disk>( | |
275 | on_disk: &'on_disk [u8], |
|
276 | on_disk: &'on_disk [u8], | |
276 | metadata: &[u8], |
|
277 | metadata: &[u8], | |
277 | ) -> Result<DirstateMap<'on_disk>, DirstateV2ParseError> { |
|
278 | ) -> Result<DirstateMap<'on_disk>, DirstateV2ParseError> { | |
278 | if on_disk.is_empty() { |
|
279 | if on_disk.is_empty() { | |
279 |
|
|
280 | let mut map = DirstateMap::empty(on_disk); | |
|
281 | map.dirstate_version = DirstateVersion::V2; | |||
|
282 | return Ok(map); | |||
280 | } |
|
283 | } | |
281 | let (meta, _) = TreeMetadata::from_bytes(metadata) |
|
284 | let (meta, _) = TreeMetadata::from_bytes(metadata) | |
282 | .map_err(|_| DirstateV2ParseError)?; |
|
285 | .map_err(|_| DirstateV2ParseError)?; | |
283 | let dirstate_map = DirstateMap { |
|
286 | let dirstate_map = DirstateMap { | |
284 | on_disk, |
|
287 | on_disk, | |
285 | root: dirstate_map::ChildNodes::OnDisk(read_nodes( |
|
288 | root: dirstate_map::ChildNodes::OnDisk(read_nodes( | |
286 | on_disk, |
|
289 | on_disk, | |
287 | meta.root_nodes, |
|
290 | meta.root_nodes, | |
288 | )?), |
|
291 | )?), | |
289 | nodes_with_entry_count: meta.nodes_with_entry_count.get(), |
|
292 | nodes_with_entry_count: meta.nodes_with_entry_count.get(), | |
290 | nodes_with_copy_source_count: meta.nodes_with_copy_source_count.get(), |
|
293 | nodes_with_copy_source_count: meta.nodes_with_copy_source_count.get(), | |
291 | ignore_patterns_hash: meta.ignore_patterns_hash, |
|
294 | ignore_patterns_hash: meta.ignore_patterns_hash, | |
292 | unreachable_bytes: meta.unreachable_bytes.get(), |
|
295 | unreachable_bytes: meta.unreachable_bytes.get(), | |
293 | old_data_size: on_disk.len(), |
|
296 | old_data_size: on_disk.len(), | |
|
297 | dirstate_version: DirstateVersion::V2, | |||
294 | }; |
|
298 | }; | |
295 | Ok(dirstate_map) |
|
299 | Ok(dirstate_map) | |
296 | } |
|
300 | } | |
297 |
|
301 | |||
298 | impl Node { |
|
302 | impl Node { | |
299 | pub(super) fn full_path<'on_disk>( |
|
303 | pub(super) fn full_path<'on_disk>( | |
300 | &self, |
|
304 | &self, | |
301 | on_disk: &'on_disk [u8], |
|
305 | on_disk: &'on_disk [u8], | |
302 | ) -> Result<&'on_disk HgPath, DirstateV2ParseError> { |
|
306 | ) -> Result<&'on_disk HgPath, DirstateV2ParseError> { | |
303 | read_hg_path(on_disk, self.full_path) |
|
307 | read_hg_path(on_disk, self.full_path) | |
304 | } |
|
308 | } | |
305 |
|
309 | |||
306 | pub(super) fn base_name_start<'on_disk>( |
|
310 | pub(super) fn base_name_start<'on_disk>( | |
307 | &self, |
|
311 | &self, | |
308 | ) -> Result<usize, DirstateV2ParseError> { |
|
312 | ) -> Result<usize, DirstateV2ParseError> { | |
309 | let start = self.base_name_start.get(); |
|
313 | let start = self.base_name_start.get(); | |
310 | if start < self.full_path.len.get() { |
|
314 | if start < self.full_path.len.get() { | |
311 | let start = usize::try_from(start) |
|
315 | let start = usize::try_from(start) | |
312 | // u32 -> usize, could only panic on a 16-bit CPU |
|
316 | // u32 -> usize, could only panic on a 16-bit CPU | |
313 | .expect("dirstate-v2 base_name_start out of bounds"); |
|
317 | .expect("dirstate-v2 base_name_start out of bounds"); | |
314 | Ok(start) |
|
318 | Ok(start) | |
315 | } else { |
|
319 | } else { | |
316 | Err(DirstateV2ParseError) |
|
320 | Err(DirstateV2ParseError) | |
317 | } |
|
321 | } | |
318 | } |
|
322 | } | |
319 |
|
323 | |||
320 | pub(super) fn base_name<'on_disk>( |
|
324 | pub(super) fn base_name<'on_disk>( | |
321 | &self, |
|
325 | &self, | |
322 | on_disk: &'on_disk [u8], |
|
326 | on_disk: &'on_disk [u8], | |
323 | ) -> Result<&'on_disk HgPath, DirstateV2ParseError> { |
|
327 | ) -> Result<&'on_disk HgPath, DirstateV2ParseError> { | |
324 | let full_path = self.full_path(on_disk)?; |
|
328 | let full_path = self.full_path(on_disk)?; | |
325 | let base_name_start = self.base_name_start()?; |
|
329 | let base_name_start = self.base_name_start()?; | |
326 | Ok(HgPath::new(&full_path.as_bytes()[base_name_start..])) |
|
330 | Ok(HgPath::new(&full_path.as_bytes()[base_name_start..])) | |
327 | } |
|
331 | } | |
328 |
|
332 | |||
329 | pub(super) fn path<'on_disk>( |
|
333 | pub(super) fn path<'on_disk>( | |
330 | &self, |
|
334 | &self, | |
331 | on_disk: &'on_disk [u8], |
|
335 | on_disk: &'on_disk [u8], | |
332 | ) -> Result<dirstate_map::NodeKey<'on_disk>, DirstateV2ParseError> { |
|
336 | ) -> Result<dirstate_map::NodeKey<'on_disk>, DirstateV2ParseError> { | |
333 | Ok(WithBasename::from_raw_parts( |
|
337 | Ok(WithBasename::from_raw_parts( | |
334 | Cow::Borrowed(self.full_path(on_disk)?), |
|
338 | Cow::Borrowed(self.full_path(on_disk)?), | |
335 | self.base_name_start()?, |
|
339 | self.base_name_start()?, | |
336 | )) |
|
340 | )) | |
337 | } |
|
341 | } | |
338 |
|
342 | |||
339 | pub(super) fn has_copy_source<'on_disk>(&self) -> bool { |
|
343 | pub(super) fn has_copy_source<'on_disk>(&self) -> bool { | |
340 | self.copy_source.start.get() != 0 |
|
344 | self.copy_source.start.get() != 0 | |
341 | } |
|
345 | } | |
342 |
|
346 | |||
343 | pub(super) fn copy_source<'on_disk>( |
|
347 | pub(super) fn copy_source<'on_disk>( | |
344 | &self, |
|
348 | &self, | |
345 | on_disk: &'on_disk [u8], |
|
349 | on_disk: &'on_disk [u8], | |
346 | ) -> Result<Option<&'on_disk HgPath>, DirstateV2ParseError> { |
|
350 | ) -> Result<Option<&'on_disk HgPath>, DirstateV2ParseError> { | |
347 | Ok(if self.has_copy_source() { |
|
351 | Ok(if self.has_copy_source() { | |
348 | Some(read_hg_path(on_disk, self.copy_source)?) |
|
352 | Some(read_hg_path(on_disk, self.copy_source)?) | |
349 | } else { |
|
353 | } else { | |
350 | None |
|
354 | None | |
351 | }) |
|
355 | }) | |
352 | } |
|
356 | } | |
353 |
|
357 | |||
354 | fn flags(&self) -> Flags { |
|
358 | fn flags(&self) -> Flags { | |
355 | Flags::from_bits_truncate(self.flags.get()) |
|
359 | Flags::from_bits_truncate(self.flags.get()) | |
356 | } |
|
360 | } | |
357 |
|
361 | |||
358 | fn has_entry(&self) -> bool { |
|
362 | fn has_entry(&self) -> bool { | |
359 | self.flags().intersects( |
|
363 | self.flags().intersects( | |
360 | Flags::WDIR_TRACKED | Flags::P1_TRACKED | Flags::P2_INFO, |
|
364 | Flags::WDIR_TRACKED | Flags::P1_TRACKED | Flags::P2_INFO, | |
361 | ) |
|
365 | ) | |
362 | } |
|
366 | } | |
363 |
|
367 | |||
364 | pub(super) fn node_data( |
|
368 | pub(super) fn node_data( | |
365 | &self, |
|
369 | &self, | |
366 | ) -> Result<dirstate_map::NodeData, DirstateV2ParseError> { |
|
370 | ) -> Result<dirstate_map::NodeData, DirstateV2ParseError> { | |
367 | if self.has_entry() { |
|
371 | if self.has_entry() { | |
368 | Ok(dirstate_map::NodeData::Entry(self.assume_entry()?)) |
|
372 | Ok(dirstate_map::NodeData::Entry(self.assume_entry()?)) | |
369 | } else if let Some(mtime) = self.cached_directory_mtime()? { |
|
373 | } else if let Some(mtime) = self.cached_directory_mtime()? { | |
370 | Ok(dirstate_map::NodeData::CachedDirectory { mtime }) |
|
374 | Ok(dirstate_map::NodeData::CachedDirectory { mtime }) | |
371 | } else { |
|
375 | } else { | |
372 | Ok(dirstate_map::NodeData::None) |
|
376 | Ok(dirstate_map::NodeData::None) | |
373 | } |
|
377 | } | |
374 | } |
|
378 | } | |
375 |
|
379 | |||
376 | pub(super) fn cached_directory_mtime( |
|
380 | pub(super) fn cached_directory_mtime( | |
377 | &self, |
|
381 | &self, | |
378 | ) -> Result<Option<TruncatedTimestamp>, DirstateV2ParseError> { |
|
382 | ) -> Result<Option<TruncatedTimestamp>, DirstateV2ParseError> { | |
379 | // For now we do not have code to handle the absence of |
|
383 | // For now we do not have code to handle the absence of | |
380 | // ALL_UNKNOWN_RECORDED, so we ignore the mtime if the flag is |
|
384 | // ALL_UNKNOWN_RECORDED, so we ignore the mtime if the flag is | |
381 | // unset. |
|
385 | // unset. | |
382 | if self.flags().contains(Flags::DIRECTORY) |
|
386 | if self.flags().contains(Flags::DIRECTORY) | |
383 | && self.flags().contains(Flags::HAS_MTIME) |
|
387 | && self.flags().contains(Flags::HAS_MTIME) | |
384 | && self.flags().contains(Flags::ALL_UNKNOWN_RECORDED) |
|
388 | && self.flags().contains(Flags::ALL_UNKNOWN_RECORDED) | |
385 | { |
|
389 | { | |
386 | Ok(Some(self.mtime()?)) |
|
390 | Ok(Some(self.mtime()?)) | |
387 | } else { |
|
391 | } else { | |
388 | Ok(None) |
|
392 | Ok(None) | |
389 | } |
|
393 | } | |
390 | } |
|
394 | } | |
391 |
|
395 | |||
392 | fn synthesize_unix_mode(&self) -> u32 { |
|
396 | fn synthesize_unix_mode(&self) -> u32 { | |
393 | let file_type = if self.flags().contains(Flags::MODE_IS_SYMLINK) { |
|
397 | let file_type = if self.flags().contains(Flags::MODE_IS_SYMLINK) { | |
394 | libc::S_IFLNK |
|
398 | libc::S_IFLNK | |
395 | } else { |
|
399 | } else { | |
396 | libc::S_IFREG |
|
400 | libc::S_IFREG | |
397 | }; |
|
401 | }; | |
398 | let permisions = if self.flags().contains(Flags::MODE_EXEC_PERM) { |
|
402 | let permisions = if self.flags().contains(Flags::MODE_EXEC_PERM) { | |
399 | 0o755 |
|
403 | 0o755 | |
400 | } else { |
|
404 | } else { | |
401 | 0o644 |
|
405 | 0o644 | |
402 | }; |
|
406 | }; | |
403 | (file_type | permisions).into() |
|
407 | (file_type | permisions).into() | |
404 | } |
|
408 | } | |
405 |
|
409 | |||
406 | fn mtime(&self) -> Result<TruncatedTimestamp, DirstateV2ParseError> { |
|
410 | fn mtime(&self) -> Result<TruncatedTimestamp, DirstateV2ParseError> { | |
407 | let mut m: TruncatedTimestamp = self.mtime.try_into()?; |
|
411 | let mut m: TruncatedTimestamp = self.mtime.try_into()?; | |
408 | if self.flags().contains(Flags::MTIME_SECOND_AMBIGUOUS) { |
|
412 | if self.flags().contains(Flags::MTIME_SECOND_AMBIGUOUS) { | |
409 | m.second_ambiguous = true; |
|
413 | m.second_ambiguous = true; | |
410 | } |
|
414 | } | |
411 | Ok(m) |
|
415 | Ok(m) | |
412 | } |
|
416 | } | |
413 |
|
417 | |||
414 | fn assume_entry(&self) -> Result<DirstateEntry, DirstateV2ParseError> { |
|
418 | fn assume_entry(&self) -> Result<DirstateEntry, DirstateV2ParseError> { | |
415 | // TODO: convert through raw bits instead? |
|
419 | // TODO: convert through raw bits instead? | |
416 | let wdir_tracked = self.flags().contains(Flags::WDIR_TRACKED); |
|
420 | let wdir_tracked = self.flags().contains(Flags::WDIR_TRACKED); | |
417 | let p1_tracked = self.flags().contains(Flags::P1_TRACKED); |
|
421 | let p1_tracked = self.flags().contains(Flags::P1_TRACKED); | |
418 | let p2_info = self.flags().contains(Flags::P2_INFO); |
|
422 | let p2_info = self.flags().contains(Flags::P2_INFO); | |
419 | let mode_size = if self.flags().contains(Flags::HAS_MODE_AND_SIZE) |
|
423 | let mode_size = if self.flags().contains(Flags::HAS_MODE_AND_SIZE) | |
420 | && !self.flags().contains(Flags::EXPECTED_STATE_IS_MODIFIED) |
|
424 | && !self.flags().contains(Flags::EXPECTED_STATE_IS_MODIFIED) | |
421 | { |
|
425 | { | |
422 | Some((self.synthesize_unix_mode(), self.size.into())) |
|
426 | Some((self.synthesize_unix_mode(), self.size.into())) | |
423 | } else { |
|
427 | } else { | |
424 | None |
|
428 | None | |
425 | }; |
|
429 | }; | |
426 | let mtime = if self.flags().contains(Flags::HAS_MTIME) |
|
430 | let mtime = if self.flags().contains(Flags::HAS_MTIME) | |
427 | && !self.flags().contains(Flags::DIRECTORY) |
|
431 | && !self.flags().contains(Flags::DIRECTORY) | |
428 | && !self.flags().contains(Flags::EXPECTED_STATE_IS_MODIFIED) |
|
432 | && !self.flags().contains(Flags::EXPECTED_STATE_IS_MODIFIED) | |
429 | { |
|
433 | { | |
430 | Some(self.mtime()?) |
|
434 | Some(self.mtime()?) | |
431 | } else { |
|
435 | } else { | |
432 | None |
|
436 | None | |
433 | }; |
|
437 | }; | |
434 | let fallback_exec = if self.flags().contains(Flags::HAS_FALLBACK_EXEC) |
|
438 | let fallback_exec = if self.flags().contains(Flags::HAS_FALLBACK_EXEC) | |
435 | { |
|
439 | { | |
436 | Some(self.flags().contains(Flags::FALLBACK_EXEC)) |
|
440 | Some(self.flags().contains(Flags::FALLBACK_EXEC)) | |
437 | } else { |
|
441 | } else { | |
438 | None |
|
442 | None | |
439 | }; |
|
443 | }; | |
440 | let fallback_symlink = |
|
444 | let fallback_symlink = | |
441 | if self.flags().contains(Flags::HAS_FALLBACK_SYMLINK) { |
|
445 | if self.flags().contains(Flags::HAS_FALLBACK_SYMLINK) { | |
442 | Some(self.flags().contains(Flags::FALLBACK_SYMLINK)) |
|
446 | Some(self.flags().contains(Flags::FALLBACK_SYMLINK)) | |
443 | } else { |
|
447 | } else { | |
444 | None |
|
448 | None | |
445 | }; |
|
449 | }; | |
446 | Ok(DirstateEntry::from_v2_data( |
|
450 | Ok(DirstateEntry::from_v2_data( | |
447 | wdir_tracked, |
|
451 | wdir_tracked, | |
448 | p1_tracked, |
|
452 | p1_tracked, | |
449 | p2_info, |
|
453 | p2_info, | |
450 | mode_size, |
|
454 | mode_size, | |
451 | mtime, |
|
455 | mtime, | |
452 | fallback_exec, |
|
456 | fallback_exec, | |
453 | fallback_symlink, |
|
457 | fallback_symlink, | |
454 | )) |
|
458 | )) | |
455 | } |
|
459 | } | |
456 |
|
460 | |||
457 | pub(super) fn entry( |
|
461 | pub(super) fn entry( | |
458 | &self, |
|
462 | &self, | |
459 | ) -> Result<Option<DirstateEntry>, DirstateV2ParseError> { |
|
463 | ) -> Result<Option<DirstateEntry>, DirstateV2ParseError> { | |
460 | if self.has_entry() { |
|
464 | if self.has_entry() { | |
461 | Ok(Some(self.assume_entry()?)) |
|
465 | Ok(Some(self.assume_entry()?)) | |
462 | } else { |
|
466 | } else { | |
463 | Ok(None) |
|
467 | Ok(None) | |
464 | } |
|
468 | } | |
465 | } |
|
469 | } | |
466 |
|
470 | |||
467 | pub(super) fn children<'on_disk>( |
|
471 | pub(super) fn children<'on_disk>( | |
468 | &self, |
|
472 | &self, | |
469 | on_disk: &'on_disk [u8], |
|
473 | on_disk: &'on_disk [u8], | |
470 | ) -> Result<&'on_disk [Node], DirstateV2ParseError> { |
|
474 | ) -> Result<&'on_disk [Node], DirstateV2ParseError> { | |
471 | read_nodes(on_disk, self.children) |
|
475 | read_nodes(on_disk, self.children) | |
472 | } |
|
476 | } | |
473 |
|
477 | |||
474 | pub(super) fn to_in_memory_node<'on_disk>( |
|
478 | pub(super) fn to_in_memory_node<'on_disk>( | |
475 | &self, |
|
479 | &self, | |
476 | on_disk: &'on_disk [u8], |
|
480 | on_disk: &'on_disk [u8], | |
477 | ) -> Result<dirstate_map::Node<'on_disk>, DirstateV2ParseError> { |
|
481 | ) -> Result<dirstate_map::Node<'on_disk>, DirstateV2ParseError> { | |
478 | Ok(dirstate_map::Node { |
|
482 | Ok(dirstate_map::Node { | |
479 | children: dirstate_map::ChildNodes::OnDisk( |
|
483 | children: dirstate_map::ChildNodes::OnDisk( | |
480 | self.children(on_disk)?, |
|
484 | self.children(on_disk)?, | |
481 | ), |
|
485 | ), | |
482 | copy_source: self.copy_source(on_disk)?.map(Cow::Borrowed), |
|
486 | copy_source: self.copy_source(on_disk)?.map(Cow::Borrowed), | |
483 | data: self.node_data()?, |
|
487 | data: self.node_data()?, | |
484 | descendants_with_entry_count: self |
|
488 | descendants_with_entry_count: self | |
485 | .descendants_with_entry_count |
|
489 | .descendants_with_entry_count | |
486 | .get(), |
|
490 | .get(), | |
487 | tracked_descendants_count: self.tracked_descendants_count.get(), |
|
491 | tracked_descendants_count: self.tracked_descendants_count.get(), | |
488 | }) |
|
492 | }) | |
489 | } |
|
493 | } | |
490 |
|
494 | |||
491 | fn from_dirstate_entry( |
|
495 | fn from_dirstate_entry( | |
492 | entry: &DirstateEntry, |
|
496 | entry: &DirstateEntry, | |
493 | ) -> (Flags, U32Be, PackedTruncatedTimestamp) { |
|
497 | ) -> (Flags, U32Be, PackedTruncatedTimestamp) { | |
494 | let ( |
|
498 | let ( | |
495 | wdir_tracked, |
|
499 | wdir_tracked, | |
496 | p1_tracked, |
|
500 | p1_tracked, | |
497 | p2_info, |
|
501 | p2_info, | |
498 | mode_size_opt, |
|
502 | mode_size_opt, | |
499 | mtime_opt, |
|
503 | mtime_opt, | |
500 | fallback_exec, |
|
504 | fallback_exec, | |
501 | fallback_symlink, |
|
505 | fallback_symlink, | |
502 | ) = entry.v2_data(); |
|
506 | ) = entry.v2_data(); | |
503 | // TODO: convert throug raw flag bits instead? |
|
507 | // TODO: convert throug raw flag bits instead? | |
504 | let mut flags = Flags::empty(); |
|
508 | let mut flags = Flags::empty(); | |
505 | flags.set(Flags::WDIR_TRACKED, wdir_tracked); |
|
509 | flags.set(Flags::WDIR_TRACKED, wdir_tracked); | |
506 | flags.set(Flags::P1_TRACKED, p1_tracked); |
|
510 | flags.set(Flags::P1_TRACKED, p1_tracked); | |
507 | flags.set(Flags::P2_INFO, p2_info); |
|
511 | flags.set(Flags::P2_INFO, p2_info); | |
508 | let size = if let Some((m, s)) = mode_size_opt { |
|
512 | let size = if let Some((m, s)) = mode_size_opt { | |
509 | let exec_perm = m & (libc::S_IXUSR as u32) != 0; |
|
513 | let exec_perm = m & (libc::S_IXUSR as u32) != 0; | |
510 | let is_symlink = m & (libc::S_IFMT as u32) == libc::S_IFLNK as u32; |
|
514 | let is_symlink = m & (libc::S_IFMT as u32) == libc::S_IFLNK as u32; | |
511 | flags.set(Flags::MODE_EXEC_PERM, exec_perm); |
|
515 | flags.set(Flags::MODE_EXEC_PERM, exec_perm); | |
512 | flags.set(Flags::MODE_IS_SYMLINK, is_symlink); |
|
516 | flags.set(Flags::MODE_IS_SYMLINK, is_symlink); | |
513 | flags.insert(Flags::HAS_MODE_AND_SIZE); |
|
517 | flags.insert(Flags::HAS_MODE_AND_SIZE); | |
514 | s.into() |
|
518 | s.into() | |
515 | } else { |
|
519 | } else { | |
516 | 0.into() |
|
520 | 0.into() | |
517 | }; |
|
521 | }; | |
518 | let mtime = if let Some(m) = mtime_opt { |
|
522 | let mtime = if let Some(m) = mtime_opt { | |
519 | flags.insert(Flags::HAS_MTIME); |
|
523 | flags.insert(Flags::HAS_MTIME); | |
520 | if m.second_ambiguous { |
|
524 | if m.second_ambiguous { | |
521 | flags.insert(Flags::MTIME_SECOND_AMBIGUOUS); |
|
525 | flags.insert(Flags::MTIME_SECOND_AMBIGUOUS); | |
522 | }; |
|
526 | }; | |
523 | m.into() |
|
527 | m.into() | |
524 | } else { |
|
528 | } else { | |
525 | PackedTruncatedTimestamp::null() |
|
529 | PackedTruncatedTimestamp::null() | |
526 | }; |
|
530 | }; | |
527 | if let Some(f_exec) = fallback_exec { |
|
531 | if let Some(f_exec) = fallback_exec { | |
528 | flags.insert(Flags::HAS_FALLBACK_EXEC); |
|
532 | flags.insert(Flags::HAS_FALLBACK_EXEC); | |
529 | if f_exec { |
|
533 | if f_exec { | |
530 | flags.insert(Flags::FALLBACK_EXEC); |
|
534 | flags.insert(Flags::FALLBACK_EXEC); | |
531 | } |
|
535 | } | |
532 | } |
|
536 | } | |
533 | if let Some(f_symlink) = fallback_symlink { |
|
537 | if let Some(f_symlink) = fallback_symlink { | |
534 | flags.insert(Flags::HAS_FALLBACK_SYMLINK); |
|
538 | flags.insert(Flags::HAS_FALLBACK_SYMLINK); | |
535 | if f_symlink { |
|
539 | if f_symlink { | |
536 | flags.insert(Flags::FALLBACK_SYMLINK); |
|
540 | flags.insert(Flags::FALLBACK_SYMLINK); | |
537 | } |
|
541 | } | |
538 | } |
|
542 | } | |
539 | (flags, size, mtime) |
|
543 | (flags, size, mtime) | |
540 | } |
|
544 | } | |
541 | } |
|
545 | } | |
542 |
|
546 | |||
543 | fn read_hg_path( |
|
547 | fn read_hg_path( | |
544 | on_disk: &[u8], |
|
548 | on_disk: &[u8], | |
545 | slice: PathSlice, |
|
549 | slice: PathSlice, | |
546 | ) -> Result<&HgPath, DirstateV2ParseError> { |
|
550 | ) -> Result<&HgPath, DirstateV2ParseError> { | |
547 | read_slice(on_disk, slice.start, slice.len.get()).map(HgPath::new) |
|
551 | read_slice(on_disk, slice.start, slice.len.get()).map(HgPath::new) | |
548 | } |
|
552 | } | |
549 |
|
553 | |||
550 | fn read_nodes( |
|
554 | fn read_nodes( | |
551 | on_disk: &[u8], |
|
555 | on_disk: &[u8], | |
552 | slice: ChildNodes, |
|
556 | slice: ChildNodes, | |
553 | ) -> Result<&[Node], DirstateV2ParseError> { |
|
557 | ) -> Result<&[Node], DirstateV2ParseError> { | |
554 | read_slice(on_disk, slice.start, slice.len.get()) |
|
558 | read_slice(on_disk, slice.start, slice.len.get()) | |
555 | } |
|
559 | } | |
556 |
|
560 | |||
557 | fn read_slice<T, Len>( |
|
561 | fn read_slice<T, Len>( | |
558 | on_disk: &[u8], |
|
562 | on_disk: &[u8], | |
559 | start: Offset, |
|
563 | start: Offset, | |
560 | len: Len, |
|
564 | len: Len, | |
561 | ) -> Result<&[T], DirstateV2ParseError> |
|
565 | ) -> Result<&[T], DirstateV2ParseError> | |
562 | where |
|
566 | where | |
563 | T: BytesCast, |
|
567 | T: BytesCast, | |
564 | Len: TryInto<usize>, |
|
568 | Len: TryInto<usize>, | |
565 | { |
|
569 | { | |
566 | // Either `usize::MAX` would result in "out of bounds" error since a single |
|
570 | // Either `usize::MAX` would result in "out of bounds" error since a single | |
567 | // `&[u8]` cannot occupy the entire addess space. |
|
571 | // `&[u8]` cannot occupy the entire addess space. | |
568 | let start = start.get().try_into().unwrap_or(std::usize::MAX); |
|
572 | let start = start.get().try_into().unwrap_or(std::usize::MAX); | |
569 | let len = len.try_into().unwrap_or(std::usize::MAX); |
|
573 | let len = len.try_into().unwrap_or(std::usize::MAX); | |
570 | on_disk |
|
574 | on_disk | |
571 | .get(start..) |
|
575 | .get(start..) | |
572 | .and_then(|bytes| T::slice_from_bytes(bytes, len).ok()) |
|
576 | .and_then(|bytes| T::slice_from_bytes(bytes, len).ok()) | |
573 | .map(|(slice, _rest)| slice) |
|
577 | .map(|(slice, _rest)| slice) | |
574 | .ok_or_else(|| DirstateV2ParseError) |
|
578 | .ok_or_else(|| DirstateV2ParseError) | |
575 | } |
|
579 | } | |
576 |
|
580 | |||
577 | pub(crate) fn for_each_tracked_path<'on_disk>( |
|
581 | pub(crate) fn for_each_tracked_path<'on_disk>( | |
578 | on_disk: &'on_disk [u8], |
|
582 | on_disk: &'on_disk [u8], | |
579 | metadata: &[u8], |
|
583 | metadata: &[u8], | |
580 | mut f: impl FnMut(&'on_disk HgPath), |
|
584 | mut f: impl FnMut(&'on_disk HgPath), | |
581 | ) -> Result<(), DirstateV2ParseError> { |
|
585 | ) -> Result<(), DirstateV2ParseError> { | |
582 | let (meta, _) = TreeMetadata::from_bytes(metadata) |
|
586 | let (meta, _) = TreeMetadata::from_bytes(metadata) | |
583 | .map_err(|_| DirstateV2ParseError)?; |
|
587 | .map_err(|_| DirstateV2ParseError)?; | |
584 | fn recur<'on_disk>( |
|
588 | fn recur<'on_disk>( | |
585 | on_disk: &'on_disk [u8], |
|
589 | on_disk: &'on_disk [u8], | |
586 | nodes: ChildNodes, |
|
590 | nodes: ChildNodes, | |
587 | f: &mut impl FnMut(&'on_disk HgPath), |
|
591 | f: &mut impl FnMut(&'on_disk HgPath), | |
588 | ) -> Result<(), DirstateV2ParseError> { |
|
592 | ) -> Result<(), DirstateV2ParseError> { | |
589 | for node in read_nodes(on_disk, nodes)? { |
|
593 | for node in read_nodes(on_disk, nodes)? { | |
590 | if let Some(entry) = node.entry()? { |
|
594 | if let Some(entry) = node.entry()? { | |
591 | if entry.state().is_tracked() { |
|
595 | if entry.state().is_tracked() { | |
592 | f(node.full_path(on_disk)?) |
|
596 | f(node.full_path(on_disk)?) | |
593 | } |
|
597 | } | |
594 | } |
|
598 | } | |
595 | recur(on_disk, node.children, f)? |
|
599 | recur(on_disk, node.children, f)? | |
596 | } |
|
600 | } | |
597 | Ok(()) |
|
601 | Ok(()) | |
598 | } |
|
602 | } | |
599 | recur(on_disk, meta.root_nodes, &mut f) |
|
603 | recur(on_disk, meta.root_nodes, &mut f) | |
600 | } |
|
604 | } | |
601 |
|
605 | |||
602 | /// Returns new data and metadata, together with whether that data should be |
|
606 | /// Returns new data and metadata, together with whether that data should be | |
603 | /// appended to the existing data file whose content is at |
|
607 | /// appended to the existing data file whose content is at | |
604 | /// `dirstate_map.on_disk` (true), instead of written to a new data file |
|
608 | /// `dirstate_map.on_disk` (true), instead of written to a new data file | |
605 | /// (false), and the previous size of data on disk. |
|
609 | /// (false), and the previous size of data on disk. | |
606 | pub(super) fn write( |
|
610 | pub(super) fn write( | |
607 | dirstate_map: &DirstateMap, |
|
611 | dirstate_map: &DirstateMap, | |
608 | can_append: bool, |
|
612 | can_append: bool, | |
609 | ) -> Result<(Vec<u8>, TreeMetadata, bool, usize), DirstateError> { |
|
613 | ) -> Result<(Vec<u8>, TreeMetadata, bool, usize), DirstateError> { | |
610 | let append = can_append && dirstate_map.write_should_append(); |
|
614 | let append = can_append && dirstate_map.write_should_append(); | |
611 |
|
615 | |||
612 | // This ignores the space for paths, and for nodes without an entry. |
|
616 | // This ignores the space for paths, and for nodes without an entry. | |
613 | // TODO: better estimate? Skip the `Vec` and write to a file directly? |
|
617 | // TODO: better estimate? Skip the `Vec` and write to a file directly? | |
614 | let size_guess = std::mem::size_of::<Node>() |
|
618 | let size_guess = std::mem::size_of::<Node>() | |
615 | * dirstate_map.nodes_with_entry_count as usize; |
|
619 | * dirstate_map.nodes_with_entry_count as usize; | |
616 |
|
620 | |||
617 | let mut writer = Writer { |
|
621 | let mut writer = Writer { | |
618 | dirstate_map, |
|
622 | dirstate_map, | |
619 | append, |
|
623 | append, | |
620 | out: Vec::with_capacity(size_guess), |
|
624 | out: Vec::with_capacity(size_guess), | |
621 | }; |
|
625 | }; | |
622 |
|
626 | |||
623 | let root_nodes = writer.write_nodes(dirstate_map.root.as_ref())?; |
|
627 | let root_nodes = writer.write_nodes(dirstate_map.root.as_ref())?; | |
624 |
|
628 | |||
625 | let unreachable_bytes = if append { |
|
629 | let unreachable_bytes = if append { | |
626 | dirstate_map.unreachable_bytes |
|
630 | dirstate_map.unreachable_bytes | |
627 | } else { |
|
631 | } else { | |
628 | 0 |
|
632 | 0 | |
629 | }; |
|
633 | }; | |
630 | let meta = TreeMetadata { |
|
634 | let meta = TreeMetadata { | |
631 | root_nodes, |
|
635 | root_nodes, | |
632 | nodes_with_entry_count: dirstate_map.nodes_with_entry_count.into(), |
|
636 | nodes_with_entry_count: dirstate_map.nodes_with_entry_count.into(), | |
633 | nodes_with_copy_source_count: dirstate_map |
|
637 | nodes_with_copy_source_count: dirstate_map | |
634 | .nodes_with_copy_source_count |
|
638 | .nodes_with_copy_source_count | |
635 | .into(), |
|
639 | .into(), | |
636 | unreachable_bytes: unreachable_bytes.into(), |
|
640 | unreachable_bytes: unreachable_bytes.into(), | |
637 | unused: [0; 4], |
|
641 | unused: [0; 4], | |
638 | ignore_patterns_hash: dirstate_map.ignore_patterns_hash, |
|
642 | ignore_patterns_hash: dirstate_map.ignore_patterns_hash, | |
639 | }; |
|
643 | }; | |
640 | Ok((writer.out, meta, append, dirstate_map.old_data_size)) |
|
644 | Ok((writer.out, meta, append, dirstate_map.old_data_size)) | |
641 | } |
|
645 | } | |
642 |
|
646 | |||
643 | struct Writer<'dmap, 'on_disk> { |
|
647 | struct Writer<'dmap, 'on_disk> { | |
644 | dirstate_map: &'dmap DirstateMap<'on_disk>, |
|
648 | dirstate_map: &'dmap DirstateMap<'on_disk>, | |
645 | append: bool, |
|
649 | append: bool, | |
646 | out: Vec<u8>, |
|
650 | out: Vec<u8>, | |
647 | } |
|
651 | } | |
648 |
|
652 | |||
649 | impl Writer<'_, '_> { |
|
653 | impl Writer<'_, '_> { | |
650 | fn write_nodes( |
|
654 | fn write_nodes( | |
651 | &mut self, |
|
655 | &mut self, | |
652 | nodes: dirstate_map::ChildNodesRef, |
|
656 | nodes: dirstate_map::ChildNodesRef, | |
653 | ) -> Result<ChildNodes, DirstateError> { |
|
657 | ) -> Result<ChildNodes, DirstateError> { | |
654 | // Reuse already-written nodes if possible |
|
658 | // Reuse already-written nodes if possible | |
655 | if self.append { |
|
659 | if self.append { | |
656 | if let dirstate_map::ChildNodesRef::OnDisk(nodes_slice) = nodes { |
|
660 | if let dirstate_map::ChildNodesRef::OnDisk(nodes_slice) = nodes { | |
657 | let start = self.on_disk_offset_of(nodes_slice).expect( |
|
661 | let start = self.on_disk_offset_of(nodes_slice).expect( | |
658 | "dirstate-v2 OnDisk nodes not found within on_disk", |
|
662 | "dirstate-v2 OnDisk nodes not found within on_disk", | |
659 | ); |
|
663 | ); | |
660 | let len = child_nodes_len_from_usize(nodes_slice.len()); |
|
664 | let len = child_nodes_len_from_usize(nodes_slice.len()); | |
661 | return Ok(ChildNodes { start, len }); |
|
665 | return Ok(ChildNodes { start, len }); | |
662 | } |
|
666 | } | |
663 | } |
|
667 | } | |
664 |
|
668 | |||
665 | // `dirstate_map::ChildNodes::InMemory` contains a `HashMap` which has |
|
669 | // `dirstate_map::ChildNodes::InMemory` contains a `HashMap` which has | |
666 | // undefined iteration order. Sort to enable binary search in the |
|
670 | // undefined iteration order. Sort to enable binary search in the | |
667 | // written file. |
|
671 | // written file. | |
668 | let nodes = nodes.sorted(); |
|
672 | let nodes = nodes.sorted(); | |
669 | let nodes_len = nodes.len(); |
|
673 | let nodes_len = nodes.len(); | |
670 |
|
674 | |||
671 | // First accumulate serialized nodes in a `Vec` |
|
675 | // First accumulate serialized nodes in a `Vec` | |
672 | let mut on_disk_nodes = Vec::with_capacity(nodes_len); |
|
676 | let mut on_disk_nodes = Vec::with_capacity(nodes_len); | |
673 | for node in nodes { |
|
677 | for node in nodes { | |
674 | let children = |
|
678 | let children = | |
675 | self.write_nodes(node.children(self.dirstate_map.on_disk)?)?; |
|
679 | self.write_nodes(node.children(self.dirstate_map.on_disk)?)?; | |
676 | let full_path = node.full_path(self.dirstate_map.on_disk)?; |
|
680 | let full_path = node.full_path(self.dirstate_map.on_disk)?; | |
677 | let full_path = self.write_path(full_path.as_bytes()); |
|
681 | let full_path = self.write_path(full_path.as_bytes()); | |
678 | let copy_source = if let Some(source) = |
|
682 | let copy_source = if let Some(source) = | |
679 | node.copy_source(self.dirstate_map.on_disk)? |
|
683 | node.copy_source(self.dirstate_map.on_disk)? | |
680 | { |
|
684 | { | |
681 | self.write_path(source.as_bytes()) |
|
685 | self.write_path(source.as_bytes()) | |
682 | } else { |
|
686 | } else { | |
683 | PathSlice { |
|
687 | PathSlice { | |
684 | start: 0.into(), |
|
688 | start: 0.into(), | |
685 | len: 0.into(), |
|
689 | len: 0.into(), | |
686 | } |
|
690 | } | |
687 | }; |
|
691 | }; | |
688 | on_disk_nodes.push(match node { |
|
692 | on_disk_nodes.push(match node { | |
689 | NodeRef::InMemory(path, node) => { |
|
693 | NodeRef::InMemory(path, node) => { | |
690 | let (flags, size, mtime) = match &node.data { |
|
694 | let (flags, size, mtime) = match &node.data { | |
691 | dirstate_map::NodeData::Entry(entry) => { |
|
695 | dirstate_map::NodeData::Entry(entry) => { | |
692 | Node::from_dirstate_entry(entry) |
|
696 | Node::from_dirstate_entry(entry) | |
693 | } |
|
697 | } | |
694 | dirstate_map::NodeData::CachedDirectory { mtime } => { |
|
698 | dirstate_map::NodeData::CachedDirectory { mtime } => { | |
695 | // we currently never set a mtime if unknown file |
|
699 | // we currently never set a mtime if unknown file | |
696 | // are present. |
|
700 | // are present. | |
697 | // So if we have a mtime for a directory, we know |
|
701 | // So if we have a mtime for a directory, we know | |
698 | // they are no unknown |
|
702 | // they are no unknown | |
699 | // files and we |
|
703 | // files and we | |
700 | // blindly set ALL_UNKNOWN_RECORDED. |
|
704 | // blindly set ALL_UNKNOWN_RECORDED. | |
701 | // |
|
705 | // | |
702 | // We never set ALL_IGNORED_RECORDED since we |
|
706 | // We never set ALL_IGNORED_RECORDED since we | |
703 | // don't track that case |
|
707 | // don't track that case | |
704 | // currently. |
|
708 | // currently. | |
705 | let mut flags = Flags::DIRECTORY |
|
709 | let mut flags = Flags::DIRECTORY | |
706 | | Flags::HAS_MTIME |
|
710 | | Flags::HAS_MTIME | |
707 | | Flags::ALL_UNKNOWN_RECORDED; |
|
711 | | Flags::ALL_UNKNOWN_RECORDED; | |
708 | if mtime.second_ambiguous { |
|
712 | if mtime.second_ambiguous { | |
709 | flags.insert(Flags::MTIME_SECOND_AMBIGUOUS) |
|
713 | flags.insert(Flags::MTIME_SECOND_AMBIGUOUS) | |
710 | } |
|
714 | } | |
711 | (flags, 0.into(), (*mtime).into()) |
|
715 | (flags, 0.into(), (*mtime).into()) | |
712 | } |
|
716 | } | |
713 | dirstate_map::NodeData::None => ( |
|
717 | dirstate_map::NodeData::None => ( | |
714 | Flags::DIRECTORY, |
|
718 | Flags::DIRECTORY, | |
715 | 0.into(), |
|
719 | 0.into(), | |
716 | PackedTruncatedTimestamp::null(), |
|
720 | PackedTruncatedTimestamp::null(), | |
717 | ), |
|
721 | ), | |
718 | }; |
|
722 | }; | |
719 | Node { |
|
723 | Node { | |
720 | children, |
|
724 | children, | |
721 | copy_source, |
|
725 | copy_source, | |
722 | full_path, |
|
726 | full_path, | |
723 | base_name_start: u16::try_from(path.base_name_start()) |
|
727 | base_name_start: u16::try_from(path.base_name_start()) | |
724 | // Could only panic for paths over 64 KiB |
|
728 | // Could only panic for paths over 64 KiB | |
725 | .expect("dirstate-v2 path length overflow") |
|
729 | .expect("dirstate-v2 path length overflow") | |
726 | .into(), |
|
730 | .into(), | |
727 | descendants_with_entry_count: node |
|
731 | descendants_with_entry_count: node | |
728 | .descendants_with_entry_count |
|
732 | .descendants_with_entry_count | |
729 | .into(), |
|
733 | .into(), | |
730 | tracked_descendants_count: node |
|
734 | tracked_descendants_count: node | |
731 | .tracked_descendants_count |
|
735 | .tracked_descendants_count | |
732 | .into(), |
|
736 | .into(), | |
733 | flags: flags.bits().into(), |
|
737 | flags: flags.bits().into(), | |
734 | size, |
|
738 | size, | |
735 | mtime, |
|
739 | mtime, | |
736 | } |
|
740 | } | |
737 | } |
|
741 | } | |
738 | NodeRef::OnDisk(node) => Node { |
|
742 | NodeRef::OnDisk(node) => Node { | |
739 | children, |
|
743 | children, | |
740 | copy_source, |
|
744 | copy_source, | |
741 | full_path, |
|
745 | full_path, | |
742 | ..*node |
|
746 | ..*node | |
743 | }, |
|
747 | }, | |
744 | }) |
|
748 | }) | |
745 | } |
|
749 | } | |
746 | // … so we can write them contiguously, after writing everything else |
|
750 | // … so we can write them contiguously, after writing everything else | |
747 | // they refer to. |
|
751 | // they refer to. | |
748 | let start = self.current_offset(); |
|
752 | let start = self.current_offset(); | |
749 | let len = child_nodes_len_from_usize(nodes_len); |
|
753 | let len = child_nodes_len_from_usize(nodes_len); | |
750 | self.out.extend(on_disk_nodes.as_bytes()); |
|
754 | self.out.extend(on_disk_nodes.as_bytes()); | |
751 | Ok(ChildNodes { start, len }) |
|
755 | Ok(ChildNodes { start, len }) | |
752 | } |
|
756 | } | |
753 |
|
757 | |||
754 | /// If the given slice of items is within `on_disk`, returns its offset |
|
758 | /// If the given slice of items is within `on_disk`, returns its offset | |
755 | /// from the start of `on_disk`. |
|
759 | /// from the start of `on_disk`. | |
756 | fn on_disk_offset_of<T>(&self, slice: &[T]) -> Option<Offset> |
|
760 | fn on_disk_offset_of<T>(&self, slice: &[T]) -> Option<Offset> | |
757 | where |
|
761 | where | |
758 | T: BytesCast, |
|
762 | T: BytesCast, | |
759 | { |
|
763 | { | |
760 | fn address_range(slice: &[u8]) -> std::ops::RangeInclusive<usize> { |
|
764 | fn address_range(slice: &[u8]) -> std::ops::RangeInclusive<usize> { | |
761 | let start = slice.as_ptr() as usize; |
|
765 | let start = slice.as_ptr() as usize; | |
762 | let end = start + slice.len(); |
|
766 | let end = start + slice.len(); | |
763 | start..=end |
|
767 | start..=end | |
764 | } |
|
768 | } | |
765 | let slice_addresses = address_range(slice.as_bytes()); |
|
769 | let slice_addresses = address_range(slice.as_bytes()); | |
766 | let on_disk_addresses = address_range(self.dirstate_map.on_disk); |
|
770 | let on_disk_addresses = address_range(self.dirstate_map.on_disk); | |
767 | if on_disk_addresses.contains(slice_addresses.start()) |
|
771 | if on_disk_addresses.contains(slice_addresses.start()) | |
768 | && on_disk_addresses.contains(slice_addresses.end()) |
|
772 | && on_disk_addresses.contains(slice_addresses.end()) | |
769 | { |
|
773 | { | |
770 | let offset = slice_addresses.start() - on_disk_addresses.start(); |
|
774 | let offset = slice_addresses.start() - on_disk_addresses.start(); | |
771 | Some(offset_from_usize(offset)) |
|
775 | Some(offset_from_usize(offset)) | |
772 | } else { |
|
776 | } else { | |
773 | None |
|
777 | None | |
774 | } |
|
778 | } | |
775 | } |
|
779 | } | |
776 |
|
780 | |||
777 | fn current_offset(&mut self) -> Offset { |
|
781 | fn current_offset(&mut self) -> Offset { | |
778 | let mut offset = self.out.len(); |
|
782 | let mut offset = self.out.len(); | |
779 | if self.append { |
|
783 | if self.append { | |
780 | offset += self.dirstate_map.on_disk.len() |
|
784 | offset += self.dirstate_map.on_disk.len() | |
781 | } |
|
785 | } | |
782 | offset_from_usize(offset) |
|
786 | offset_from_usize(offset) | |
783 | } |
|
787 | } | |
784 |
|
788 | |||
785 | fn write_path(&mut self, slice: &[u8]) -> PathSlice { |
|
789 | fn write_path(&mut self, slice: &[u8]) -> PathSlice { | |
786 | let len = path_len_from_usize(slice.len()); |
|
790 | let len = path_len_from_usize(slice.len()); | |
787 | // Reuse an already-written path if possible |
|
791 | // Reuse an already-written path if possible | |
788 | if self.append { |
|
792 | if self.append { | |
789 | if let Some(start) = self.on_disk_offset_of(slice) { |
|
793 | if let Some(start) = self.on_disk_offset_of(slice) { | |
790 | return PathSlice { start, len }; |
|
794 | return PathSlice { start, len }; | |
791 | } |
|
795 | } | |
792 | } |
|
796 | } | |
793 | let start = self.current_offset(); |
|
797 | let start = self.current_offset(); | |
794 | self.out.extend(slice.as_bytes()); |
|
798 | self.out.extend(slice.as_bytes()); | |
795 | PathSlice { start, len } |
|
799 | PathSlice { start, len } | |
796 | } |
|
800 | } | |
797 | } |
|
801 | } | |
798 |
|
802 | |||
799 | fn offset_from_usize(x: usize) -> Offset { |
|
803 | fn offset_from_usize(x: usize) -> Offset { | |
800 | u32::try_from(x) |
|
804 | u32::try_from(x) | |
801 | // Could only panic for a dirstate file larger than 4 GiB |
|
805 | // Could only panic for a dirstate file larger than 4 GiB | |
802 | .expect("dirstate-v2 offset overflow") |
|
806 | .expect("dirstate-v2 offset overflow") | |
803 | .into() |
|
807 | .into() | |
804 | } |
|
808 | } | |
805 |
|
809 | |||
806 | fn child_nodes_len_from_usize(x: usize) -> Size { |
|
810 | fn child_nodes_len_from_usize(x: usize) -> Size { | |
807 | u32::try_from(x) |
|
811 | u32::try_from(x) | |
808 | // Could only panic with over 4 billion nodes |
|
812 | // Could only panic with over 4 billion nodes | |
809 | .expect("dirstate-v2 slice length overflow") |
|
813 | .expect("dirstate-v2 slice length overflow") | |
810 | .into() |
|
814 | .into() | |
811 | } |
|
815 | } | |
812 |
|
816 | |||
813 | fn path_len_from_usize(x: usize) -> PathSize { |
|
817 | fn path_len_from_usize(x: usize) -> PathSize { | |
814 | u16::try_from(x) |
|
818 | u16::try_from(x) | |
815 | // Could only panic for paths over 64 KiB |
|
819 | // Could only panic for paths over 64 KiB | |
816 | .expect("dirstate-v2 path length overflow") |
|
820 | .expect("dirstate-v2 path length overflow") | |
817 | .into() |
|
821 | .into() | |
818 | } |
|
822 | } | |
819 |
|
823 | |||
820 | impl From<TruncatedTimestamp> for PackedTruncatedTimestamp { |
|
824 | impl From<TruncatedTimestamp> for PackedTruncatedTimestamp { | |
821 | fn from(timestamp: TruncatedTimestamp) -> Self { |
|
825 | fn from(timestamp: TruncatedTimestamp) -> Self { | |
822 | Self { |
|
826 | Self { | |
823 | truncated_seconds: timestamp.truncated_seconds().into(), |
|
827 | truncated_seconds: timestamp.truncated_seconds().into(), | |
824 | nanoseconds: timestamp.nanoseconds().into(), |
|
828 | nanoseconds: timestamp.nanoseconds().into(), | |
825 | } |
|
829 | } | |
826 | } |
|
830 | } | |
827 | } |
|
831 | } | |
828 |
|
832 | |||
829 | impl TryFrom<PackedTruncatedTimestamp> for TruncatedTimestamp { |
|
833 | impl TryFrom<PackedTruncatedTimestamp> for TruncatedTimestamp { | |
830 | type Error = DirstateV2ParseError; |
|
834 | type Error = DirstateV2ParseError; | |
831 |
|
835 | |||
832 | fn try_from( |
|
836 | fn try_from( | |
833 | timestamp: PackedTruncatedTimestamp, |
|
837 | timestamp: PackedTruncatedTimestamp, | |
834 | ) -> Result<Self, Self::Error> { |
|
838 | ) -> Result<Self, Self::Error> { | |
835 | Self::from_already_truncated( |
|
839 | Self::from_already_truncated( | |
836 | timestamp.truncated_seconds.get(), |
|
840 | timestamp.truncated_seconds.get(), | |
837 | timestamp.nanoseconds.get(), |
|
841 | timestamp.nanoseconds.get(), | |
838 | false, |
|
842 | false, | |
839 | ) |
|
843 | ) | |
840 | } |
|
844 | } | |
841 | } |
|
845 | } | |
842 | impl PackedTruncatedTimestamp { |
|
846 | impl PackedTruncatedTimestamp { | |
843 | fn null() -> Self { |
|
847 | fn null() -> Self { | |
844 | Self { |
|
848 | Self { | |
845 | truncated_seconds: 0.into(), |
|
849 | truncated_seconds: 0.into(), | |
846 | nanoseconds: 0.into(), |
|
850 | nanoseconds: 0.into(), | |
847 | } |
|
851 | } | |
848 | } |
|
852 | } | |
849 | } |
|
853 | } |
@@ -1,849 +1,864 b'' | |||||
1 | use crate::dirstate::entry::TruncatedTimestamp; |
|
1 | use crate::dirstate::entry::TruncatedTimestamp; | |
2 | use crate::dirstate::status::IgnoreFnType; |
|
2 | use crate::dirstate::status::IgnoreFnType; | |
3 | use crate::dirstate::status::StatusPath; |
|
3 | use crate::dirstate::status::StatusPath; | |
4 | use crate::dirstate_tree::dirstate_map::BorrowedPath; |
|
4 | use crate::dirstate_tree::dirstate_map::BorrowedPath; | |
5 | use crate::dirstate_tree::dirstate_map::ChildNodesRef; |
|
5 | use crate::dirstate_tree::dirstate_map::ChildNodesRef; | |
6 | use crate::dirstate_tree::dirstate_map::DirstateMap; |
|
6 | use crate::dirstate_tree::dirstate_map::DirstateMap; | |
|
7 | use crate::dirstate_tree::dirstate_map::DirstateVersion; | |||
7 | use crate::dirstate_tree::dirstate_map::NodeData; |
|
8 | use crate::dirstate_tree::dirstate_map::NodeData; | |
8 | use crate::dirstate_tree::dirstate_map::NodeRef; |
|
9 | use crate::dirstate_tree::dirstate_map::NodeRef; | |
9 | use crate::dirstate_tree::on_disk::DirstateV2ParseError; |
|
10 | use crate::dirstate_tree::on_disk::DirstateV2ParseError; | |
10 | use crate::matchers::get_ignore_function; |
|
11 | use crate::matchers::get_ignore_function; | |
11 | use crate::matchers::Matcher; |
|
12 | use crate::matchers::Matcher; | |
12 | use crate::utils::files::get_bytes_from_os_string; |
|
13 | use crate::utils::files::get_bytes_from_os_string; | |
13 | use crate::utils::files::get_path_from_bytes; |
|
14 | use crate::utils::files::get_path_from_bytes; | |
14 | use crate::utils::hg_path::HgPath; |
|
15 | use crate::utils::hg_path::HgPath; | |
15 | use crate::BadMatch; |
|
16 | use crate::BadMatch; | |
16 | use crate::DirstateStatus; |
|
17 | use crate::DirstateStatus; | |
17 | use crate::EntryState; |
|
18 | use crate::EntryState; | |
18 | use crate::HgPathBuf; |
|
19 | use crate::HgPathBuf; | |
19 | use crate::HgPathCow; |
|
20 | use crate::HgPathCow; | |
20 | use crate::PatternFileWarning; |
|
21 | use crate::PatternFileWarning; | |
21 | use crate::StatusError; |
|
22 | use crate::StatusError; | |
22 | use crate::StatusOptions; |
|
23 | use crate::StatusOptions; | |
23 | use micro_timer::timed; |
|
24 | use micro_timer::timed; | |
24 | use rayon::prelude::*; |
|
25 | use rayon::prelude::*; | |
25 | use sha1::{Digest, Sha1}; |
|
26 | use sha1::{Digest, Sha1}; | |
26 | use std::borrow::Cow; |
|
27 | use std::borrow::Cow; | |
27 | use std::io; |
|
28 | use std::io; | |
28 | use std::path::Path; |
|
29 | use std::path::Path; | |
29 | use std::path::PathBuf; |
|
30 | use std::path::PathBuf; | |
30 | use std::sync::Mutex; |
|
31 | use std::sync::Mutex; | |
31 | use std::time::SystemTime; |
|
32 | use std::time::SystemTime; | |
32 |
|
33 | |||
33 | /// Returns the status of the working directory compared to its parent |
|
34 | /// Returns the status of the working directory compared to its parent | |
34 | /// changeset. |
|
35 | /// changeset. | |
35 | /// |
|
36 | /// | |
36 | /// This algorithm is based on traversing the filesystem tree (`fs` in function |
|
37 | /// This algorithm is based on traversing the filesystem tree (`fs` in function | |
37 | /// and variable names) and dirstate tree at the same time. The core of this |
|
38 | /// and variable names) and dirstate tree at the same time. The core of this | |
38 | /// traversal is the recursive `traverse_fs_directory_and_dirstate` function |
|
39 | /// traversal is the recursive `traverse_fs_directory_and_dirstate` function | |
39 | /// and its use of `itertools::merge_join_by`. When reaching a path that only |
|
40 | /// and its use of `itertools::merge_join_by`. When reaching a path that only | |
40 | /// exists in one of the two trees, depending on information requested by |
|
41 | /// exists in one of the two trees, depending on information requested by | |
41 | /// `options` we may need to traverse the remaining subtree. |
|
42 | /// `options` we may need to traverse the remaining subtree. | |
42 | #[timed] |
|
43 | #[timed] | |
43 | pub fn status<'dirstate>( |
|
44 | pub fn status<'dirstate>( | |
44 | dmap: &'dirstate mut DirstateMap, |
|
45 | dmap: &'dirstate mut DirstateMap, | |
45 | matcher: &(dyn Matcher + Sync), |
|
46 | matcher: &(dyn Matcher + Sync), | |
46 | root_dir: PathBuf, |
|
47 | root_dir: PathBuf, | |
47 | ignore_files: Vec<PathBuf>, |
|
48 | ignore_files: Vec<PathBuf>, | |
48 | options: StatusOptions, |
|
49 | options: StatusOptions, | |
49 | ) -> Result<(DirstateStatus<'dirstate>, Vec<PatternFileWarning>), StatusError> |
|
50 | ) -> Result<(DirstateStatus<'dirstate>, Vec<PatternFileWarning>), StatusError> | |
50 | { |
|
51 | { | |
51 | // Force the global rayon threadpool to not exceed 16 concurrent threads. |
|
52 | // Force the global rayon threadpool to not exceed 16 concurrent threads. | |
52 | // This is a stop-gap measure until we figure out why using more than 16 |
|
53 | // This is a stop-gap measure until we figure out why using more than 16 | |
53 | // threads makes `status` slower for each additional thread. |
|
54 | // threads makes `status` slower for each additional thread. | |
54 | // We use `ok()` in case the global threadpool has already been |
|
55 | // We use `ok()` in case the global threadpool has already been | |
55 | // instantiated in `rhg` or some other caller. |
|
56 | // instantiated in `rhg` or some other caller. | |
56 | // TODO find the underlying cause and fix it, then remove this. |
|
57 | // TODO find the underlying cause and fix it, then remove this. | |
57 | rayon::ThreadPoolBuilder::new() |
|
58 | rayon::ThreadPoolBuilder::new() | |
58 | .num_threads(16) |
|
59 | .num_threads(16) | |
59 | .build_global() |
|
60 | .build_global() | |
60 | .ok(); |
|
61 | .ok(); | |
61 |
|
62 | |||
62 | let (ignore_fn, warnings, patterns_changed): (IgnoreFnType, _, _) = |
|
63 | let (ignore_fn, warnings, patterns_changed): (IgnoreFnType, _, _) = | |
63 | if options.list_ignored || options.list_unknown { |
|
64 | if options.list_ignored || options.list_unknown { | |
64 | let mut hasher = Sha1::new(); |
|
65 | let (ignore_fn, warnings, changed) = match dmap.dirstate_version { | |
65 | let (ignore_fn, warnings) = get_ignore_function( |
|
66 | DirstateVersion::V1 => { | |
66 | ignore_files, |
|
67 | let (ignore_fn, warnings) = get_ignore_function( | |
67 |
|
|
68 | ignore_files, | |
68 | &mut |pattern_bytes| hasher.update(pattern_bytes), |
|
69 | &root_dir, | |
69 | )?; |
|
70 | &mut |_pattern_bytes| {}, | |
70 | let new_hash = *hasher.finalize().as_ref(); |
|
71 | )?; | |
71 | let changed = new_hash != dmap.ignore_patterns_hash; |
|
72 | (ignore_fn, warnings, None) | |
72 | dmap.ignore_patterns_hash = new_hash; |
|
73 | } | |
73 | (ignore_fn, warnings, Some(changed)) |
|
74 | DirstateVersion::V2 => { | |
|
75 | let mut hasher = Sha1::new(); | |||
|
76 | let (ignore_fn, warnings) = get_ignore_function( | |||
|
77 | ignore_files, | |||
|
78 | &root_dir, | |||
|
79 | &mut |pattern_bytes| hasher.update(pattern_bytes), | |||
|
80 | )?; | |||
|
81 | let new_hash = *hasher.finalize().as_ref(); | |||
|
82 | let changed = new_hash != dmap.ignore_patterns_hash; | |||
|
83 | dmap.ignore_patterns_hash = new_hash; | |||
|
84 | (ignore_fn, warnings, Some(changed)) | |||
|
85 | } | |||
|
86 | }; | |||
|
87 | (ignore_fn, warnings, changed) | |||
74 | } else { |
|
88 | } else { | |
75 | (Box::new(|&_| true), vec![], None) |
|
89 | (Box::new(|&_| true), vec![], None) | |
76 | }; |
|
90 | }; | |
77 |
|
91 | |||
78 | let filesystem_time_at_status_start = |
|
92 | let filesystem_time_at_status_start = | |
79 | filesystem_now(&root_dir).ok().map(TruncatedTimestamp::from); |
|
93 | filesystem_now(&root_dir).ok().map(TruncatedTimestamp::from); | |
80 |
|
94 | |||
81 | // If the repository is under the current directory, prefer using a |
|
95 | // If the repository is under the current directory, prefer using a | |
82 | // relative path, so the kernel needs to traverse fewer directory in every |
|
96 | // relative path, so the kernel needs to traverse fewer directory in every | |
83 | // call to `read_dir` or `symlink_metadata`. |
|
97 | // call to `read_dir` or `symlink_metadata`. | |
84 | // This is effective in the common case where the current directory is the |
|
98 | // This is effective in the common case where the current directory is the | |
85 | // repository root. |
|
99 | // repository root. | |
86 |
|
100 | |||
87 | // TODO: Better yet would be to use libc functions like `openat` and |
|
101 | // TODO: Better yet would be to use libc functions like `openat` and | |
88 | // `fstatat` to remove such repeated traversals entirely, but the standard |
|
102 | // `fstatat` to remove such repeated traversals entirely, but the standard | |
89 | // library does not provide APIs based on those. |
|
103 | // library does not provide APIs based on those. | |
90 | // Maybe with a crate like https://crates.io/crates/openat instead? |
|
104 | // Maybe with a crate like https://crates.io/crates/openat instead? | |
91 | let root_dir = if let Some(relative) = std::env::current_dir() |
|
105 | let root_dir = if let Some(relative) = std::env::current_dir() | |
92 | .ok() |
|
106 | .ok() | |
93 | .and_then(|cwd| root_dir.strip_prefix(cwd).ok()) |
|
107 | .and_then(|cwd| root_dir.strip_prefix(cwd).ok()) | |
94 | { |
|
108 | { | |
95 | relative |
|
109 | relative | |
96 | } else { |
|
110 | } else { | |
97 | &root_dir |
|
111 | &root_dir | |
98 | }; |
|
112 | }; | |
99 |
|
113 | |||
100 | let outcome = DirstateStatus { |
|
114 | let outcome = DirstateStatus { | |
101 | filesystem_time_at_status_start, |
|
115 | filesystem_time_at_status_start, | |
102 | ..Default::default() |
|
116 | ..Default::default() | |
103 | }; |
|
117 | }; | |
104 | let common = StatusCommon { |
|
118 | let common = StatusCommon { | |
105 | dmap, |
|
119 | dmap, | |
106 | options, |
|
120 | options, | |
107 | matcher, |
|
121 | matcher, | |
108 | ignore_fn, |
|
122 | ignore_fn, | |
109 | outcome: Mutex::new(outcome), |
|
123 | outcome: Mutex::new(outcome), | |
110 | ignore_patterns_have_changed: patterns_changed, |
|
124 | ignore_patterns_have_changed: patterns_changed, | |
111 | new_cachable_directories: Default::default(), |
|
125 | new_cachable_directories: Default::default(), | |
112 | outated_cached_directories: Default::default(), |
|
126 | outated_cached_directories: Default::default(), | |
113 | filesystem_time_at_status_start, |
|
127 | filesystem_time_at_status_start, | |
114 | }; |
|
128 | }; | |
115 | let is_at_repo_root = true; |
|
129 | let is_at_repo_root = true; | |
116 | let hg_path = &BorrowedPath::OnDisk(HgPath::new("")); |
|
130 | let hg_path = &BorrowedPath::OnDisk(HgPath::new("")); | |
117 | let has_ignored_ancestor = false; |
|
131 | let has_ignored_ancestor = false; | |
118 | let root_cached_mtime = None; |
|
132 | let root_cached_mtime = None; | |
119 | let root_dir_metadata = None; |
|
133 | let root_dir_metadata = None; | |
120 | // If the path we have for the repository root is a symlink, do follow it. |
|
134 | // If the path we have for the repository root is a symlink, do follow it. | |
121 | // (As opposed to symlinks within the working directory which are not |
|
135 | // (As opposed to symlinks within the working directory which are not | |
122 | // followed, using `std::fs::symlink_metadata`.) |
|
136 | // followed, using `std::fs::symlink_metadata`.) | |
123 | common.traverse_fs_directory_and_dirstate( |
|
137 | common.traverse_fs_directory_and_dirstate( | |
124 | has_ignored_ancestor, |
|
138 | has_ignored_ancestor, | |
125 | dmap.root.as_ref(), |
|
139 | dmap.root.as_ref(), | |
126 | hg_path, |
|
140 | hg_path, | |
127 | &root_dir, |
|
141 | &root_dir, | |
128 | root_dir_metadata, |
|
142 | root_dir_metadata, | |
129 | root_cached_mtime, |
|
143 | root_cached_mtime, | |
130 | is_at_repo_root, |
|
144 | is_at_repo_root, | |
131 | )?; |
|
145 | )?; | |
132 | let mut outcome = common.outcome.into_inner().unwrap(); |
|
146 | let mut outcome = common.outcome.into_inner().unwrap(); | |
133 | let new_cachable = common.new_cachable_directories.into_inner().unwrap(); |
|
147 | let new_cachable = common.new_cachable_directories.into_inner().unwrap(); | |
134 | let outdated = common.outated_cached_directories.into_inner().unwrap(); |
|
148 | let outdated = common.outated_cached_directories.into_inner().unwrap(); | |
135 |
|
149 | |||
136 | outcome.dirty = common.ignore_patterns_have_changed == Some(true) |
|
150 | outcome.dirty = common.ignore_patterns_have_changed == Some(true) | |
137 | || !outdated.is_empty() |
|
151 | || !outdated.is_empty() | |
138 |
|| !new_cachable.is_empty() |
|
152 | || (!new_cachable.is_empty() | |
|
153 | && dmap.dirstate_version == DirstateVersion::V2); | |||
139 |
|
154 | |||
140 | // Remove outdated mtimes before adding new mtimes, in case a given |
|
155 | // Remove outdated mtimes before adding new mtimes, in case a given | |
141 | // directory is both |
|
156 | // directory is both | |
142 | for path in &outdated { |
|
157 | for path in &outdated { | |
143 | let node = dmap.get_or_insert(path)?; |
|
158 | let node = dmap.get_or_insert(path)?; | |
144 | if let NodeData::CachedDirectory { .. } = &node.data { |
|
159 | if let NodeData::CachedDirectory { .. } = &node.data { | |
145 | node.data = NodeData::None |
|
160 | node.data = NodeData::None | |
146 | } |
|
161 | } | |
147 | } |
|
162 | } | |
148 | for (path, mtime) in &new_cachable { |
|
163 | for (path, mtime) in &new_cachable { | |
149 | let node = dmap.get_or_insert(path)?; |
|
164 | let node = dmap.get_or_insert(path)?; | |
150 | match &node.data { |
|
165 | match &node.data { | |
151 | NodeData::Entry(_) => {} // Don’t overwrite an entry |
|
166 | NodeData::Entry(_) => {} // Don’t overwrite an entry | |
152 | NodeData::CachedDirectory { .. } | NodeData::None => { |
|
167 | NodeData::CachedDirectory { .. } | NodeData::None => { | |
153 | node.data = NodeData::CachedDirectory { mtime: *mtime } |
|
168 | node.data = NodeData::CachedDirectory { mtime: *mtime } | |
154 | } |
|
169 | } | |
155 | } |
|
170 | } | |
156 | } |
|
171 | } | |
157 |
|
172 | |||
158 | Ok((outcome, warnings)) |
|
173 | Ok((outcome, warnings)) | |
159 | } |
|
174 | } | |
160 |
|
175 | |||
161 | /// Bag of random things needed by various parts of the algorithm. Reduces the |
|
176 | /// Bag of random things needed by various parts of the algorithm. Reduces the | |
162 | /// number of parameters passed to functions. |
|
177 | /// number of parameters passed to functions. | |
163 | struct StatusCommon<'a, 'tree, 'on_disk: 'tree> { |
|
178 | struct StatusCommon<'a, 'tree, 'on_disk: 'tree> { | |
164 | dmap: &'tree DirstateMap<'on_disk>, |
|
179 | dmap: &'tree DirstateMap<'on_disk>, | |
165 | options: StatusOptions, |
|
180 | options: StatusOptions, | |
166 | matcher: &'a (dyn Matcher + Sync), |
|
181 | matcher: &'a (dyn Matcher + Sync), | |
167 | ignore_fn: IgnoreFnType<'a>, |
|
182 | ignore_fn: IgnoreFnType<'a>, | |
168 | outcome: Mutex<DirstateStatus<'on_disk>>, |
|
183 | outcome: Mutex<DirstateStatus<'on_disk>>, | |
169 | new_cachable_directories: |
|
184 | new_cachable_directories: | |
170 | Mutex<Vec<(Cow<'on_disk, HgPath>, TruncatedTimestamp)>>, |
|
185 | Mutex<Vec<(Cow<'on_disk, HgPath>, TruncatedTimestamp)>>, | |
171 | outated_cached_directories: Mutex<Vec<Cow<'on_disk, HgPath>>>, |
|
186 | outated_cached_directories: Mutex<Vec<Cow<'on_disk, HgPath>>>, | |
172 |
|
187 | |||
173 | /// Whether ignore files like `.hgignore` have changed since the previous |
|
188 | /// Whether ignore files like `.hgignore` have changed since the previous | |
174 | /// time a `status()` call wrote their hash to the dirstate. `None` means |
|
189 | /// time a `status()` call wrote their hash to the dirstate. `None` means | |
175 | /// we don’t know as this run doesn’t list either ignored or uknown files |
|
190 | /// we don’t know as this run doesn’t list either ignored or uknown files | |
176 | /// and therefore isn’t reading `.hgignore`. |
|
191 | /// and therefore isn’t reading `.hgignore`. | |
177 | ignore_patterns_have_changed: Option<bool>, |
|
192 | ignore_patterns_have_changed: Option<bool>, | |
178 |
|
193 | |||
179 | /// The current time at the start of the `status()` algorithm, as measured |
|
194 | /// The current time at the start of the `status()` algorithm, as measured | |
180 | /// and possibly truncated by the filesystem. |
|
195 | /// and possibly truncated by the filesystem. | |
181 | filesystem_time_at_status_start: Option<TruncatedTimestamp>, |
|
196 | filesystem_time_at_status_start: Option<TruncatedTimestamp>, | |
182 | } |
|
197 | } | |
183 |
|
198 | |||
184 | enum Outcome { |
|
199 | enum Outcome { | |
185 | Modified, |
|
200 | Modified, | |
186 | Added, |
|
201 | Added, | |
187 | Removed, |
|
202 | Removed, | |
188 | Deleted, |
|
203 | Deleted, | |
189 | Clean, |
|
204 | Clean, | |
190 | Ignored, |
|
205 | Ignored, | |
191 | Unknown, |
|
206 | Unknown, | |
192 | Unsure, |
|
207 | Unsure, | |
193 | } |
|
208 | } | |
194 |
|
209 | |||
195 | impl<'a, 'tree, 'on_disk> StatusCommon<'a, 'tree, 'on_disk> { |
|
210 | impl<'a, 'tree, 'on_disk> StatusCommon<'a, 'tree, 'on_disk> { | |
196 | fn push_outcome( |
|
211 | fn push_outcome( | |
197 | &self, |
|
212 | &self, | |
198 | which: Outcome, |
|
213 | which: Outcome, | |
199 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
214 | dirstate_node: &NodeRef<'tree, 'on_disk>, | |
200 | ) -> Result<(), DirstateV2ParseError> { |
|
215 | ) -> Result<(), DirstateV2ParseError> { | |
201 | let path = dirstate_node |
|
216 | let path = dirstate_node | |
202 | .full_path_borrowed(self.dmap.on_disk)? |
|
217 | .full_path_borrowed(self.dmap.on_disk)? | |
203 | .detach_from_tree(); |
|
218 | .detach_from_tree(); | |
204 | let copy_source = if self.options.list_copies { |
|
219 | let copy_source = if self.options.list_copies { | |
205 | dirstate_node |
|
220 | dirstate_node | |
206 | .copy_source_borrowed(self.dmap.on_disk)? |
|
221 | .copy_source_borrowed(self.dmap.on_disk)? | |
207 | .map(|source| source.detach_from_tree()) |
|
222 | .map(|source| source.detach_from_tree()) | |
208 | } else { |
|
223 | } else { | |
209 | None |
|
224 | None | |
210 | }; |
|
225 | }; | |
211 | self.push_outcome_common(which, path, copy_source); |
|
226 | self.push_outcome_common(which, path, copy_source); | |
212 | Ok(()) |
|
227 | Ok(()) | |
213 | } |
|
228 | } | |
214 |
|
229 | |||
215 | fn push_outcome_without_copy_source( |
|
230 | fn push_outcome_without_copy_source( | |
216 | &self, |
|
231 | &self, | |
217 | which: Outcome, |
|
232 | which: Outcome, | |
218 | path: &BorrowedPath<'_, 'on_disk>, |
|
233 | path: &BorrowedPath<'_, 'on_disk>, | |
219 | ) { |
|
234 | ) { | |
220 | self.push_outcome_common(which, path.detach_from_tree(), None) |
|
235 | self.push_outcome_common(which, path.detach_from_tree(), None) | |
221 | } |
|
236 | } | |
222 |
|
237 | |||
223 | fn push_outcome_common( |
|
238 | fn push_outcome_common( | |
224 | &self, |
|
239 | &self, | |
225 | which: Outcome, |
|
240 | which: Outcome, | |
226 | path: HgPathCow<'on_disk>, |
|
241 | path: HgPathCow<'on_disk>, | |
227 | copy_source: Option<HgPathCow<'on_disk>>, |
|
242 | copy_source: Option<HgPathCow<'on_disk>>, | |
228 | ) { |
|
243 | ) { | |
229 | let mut outcome = self.outcome.lock().unwrap(); |
|
244 | let mut outcome = self.outcome.lock().unwrap(); | |
230 | let vec = match which { |
|
245 | let vec = match which { | |
231 | Outcome::Modified => &mut outcome.modified, |
|
246 | Outcome::Modified => &mut outcome.modified, | |
232 | Outcome::Added => &mut outcome.added, |
|
247 | Outcome::Added => &mut outcome.added, | |
233 | Outcome::Removed => &mut outcome.removed, |
|
248 | Outcome::Removed => &mut outcome.removed, | |
234 | Outcome::Deleted => &mut outcome.deleted, |
|
249 | Outcome::Deleted => &mut outcome.deleted, | |
235 | Outcome::Clean => &mut outcome.clean, |
|
250 | Outcome::Clean => &mut outcome.clean, | |
236 | Outcome::Ignored => &mut outcome.ignored, |
|
251 | Outcome::Ignored => &mut outcome.ignored, | |
237 | Outcome::Unknown => &mut outcome.unknown, |
|
252 | Outcome::Unknown => &mut outcome.unknown, | |
238 | Outcome::Unsure => &mut outcome.unsure, |
|
253 | Outcome::Unsure => &mut outcome.unsure, | |
239 | }; |
|
254 | }; | |
240 | vec.push(StatusPath { path, copy_source }); |
|
255 | vec.push(StatusPath { path, copy_source }); | |
241 | } |
|
256 | } | |
242 |
|
257 | |||
243 | fn read_dir( |
|
258 | fn read_dir( | |
244 | &self, |
|
259 | &self, | |
245 | hg_path: &HgPath, |
|
260 | hg_path: &HgPath, | |
246 | fs_path: &Path, |
|
261 | fs_path: &Path, | |
247 | is_at_repo_root: bool, |
|
262 | is_at_repo_root: bool, | |
248 | ) -> Result<Vec<DirEntry>, ()> { |
|
263 | ) -> Result<Vec<DirEntry>, ()> { | |
249 | DirEntry::read_dir(fs_path, is_at_repo_root) |
|
264 | DirEntry::read_dir(fs_path, is_at_repo_root) | |
250 | .map_err(|error| self.io_error(error, hg_path)) |
|
265 | .map_err(|error| self.io_error(error, hg_path)) | |
251 | } |
|
266 | } | |
252 |
|
267 | |||
253 | fn io_error(&self, error: std::io::Error, hg_path: &HgPath) { |
|
268 | fn io_error(&self, error: std::io::Error, hg_path: &HgPath) { | |
254 | let errno = error.raw_os_error().expect("expected real OS error"); |
|
269 | let errno = error.raw_os_error().expect("expected real OS error"); | |
255 | self.outcome |
|
270 | self.outcome | |
256 | .lock() |
|
271 | .lock() | |
257 | .unwrap() |
|
272 | .unwrap() | |
258 | .bad |
|
273 | .bad | |
259 | .push((hg_path.to_owned().into(), BadMatch::OsError(errno))) |
|
274 | .push((hg_path.to_owned().into(), BadMatch::OsError(errno))) | |
260 | } |
|
275 | } | |
261 |
|
276 | |||
262 | fn check_for_outdated_directory_cache( |
|
277 | fn check_for_outdated_directory_cache( | |
263 | &self, |
|
278 | &self, | |
264 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
279 | dirstate_node: &NodeRef<'tree, 'on_disk>, | |
265 | ) -> Result<(), DirstateV2ParseError> { |
|
280 | ) -> Result<(), DirstateV2ParseError> { | |
266 | if self.ignore_patterns_have_changed == Some(true) |
|
281 | if self.ignore_patterns_have_changed == Some(true) | |
267 | && dirstate_node.cached_directory_mtime()?.is_some() |
|
282 | && dirstate_node.cached_directory_mtime()?.is_some() | |
268 | { |
|
283 | { | |
269 | self.outated_cached_directories.lock().unwrap().push( |
|
284 | self.outated_cached_directories.lock().unwrap().push( | |
270 | dirstate_node |
|
285 | dirstate_node | |
271 | .full_path_borrowed(self.dmap.on_disk)? |
|
286 | .full_path_borrowed(self.dmap.on_disk)? | |
272 | .detach_from_tree(), |
|
287 | .detach_from_tree(), | |
273 | ) |
|
288 | ) | |
274 | } |
|
289 | } | |
275 | Ok(()) |
|
290 | Ok(()) | |
276 | } |
|
291 | } | |
277 |
|
292 | |||
278 | /// If this returns true, we can get accurate results by only using |
|
293 | /// If this returns true, we can get accurate results by only using | |
279 | /// `symlink_metadata` for child nodes that exist in the dirstate and don’t |
|
294 | /// `symlink_metadata` for child nodes that exist in the dirstate and don’t | |
280 | /// need to call `read_dir`. |
|
295 | /// need to call `read_dir`. | |
281 | fn can_skip_fs_readdir( |
|
296 | fn can_skip_fs_readdir( | |
282 | &self, |
|
297 | &self, | |
283 | directory_metadata: Option<&std::fs::Metadata>, |
|
298 | directory_metadata: Option<&std::fs::Metadata>, | |
284 | cached_directory_mtime: Option<TruncatedTimestamp>, |
|
299 | cached_directory_mtime: Option<TruncatedTimestamp>, | |
285 | ) -> bool { |
|
300 | ) -> bool { | |
286 | if !self.options.list_unknown && !self.options.list_ignored { |
|
301 | if !self.options.list_unknown && !self.options.list_ignored { | |
287 | // All states that we care about listing have corresponding |
|
302 | // All states that we care about listing have corresponding | |
288 | // dirstate entries. |
|
303 | // dirstate entries. | |
289 | // This happens for example with `hg status -mard`. |
|
304 | // This happens for example with `hg status -mard`. | |
290 | return true; |
|
305 | return true; | |
291 | } |
|
306 | } | |
292 | if !self.options.list_ignored |
|
307 | if !self.options.list_ignored | |
293 | && self.ignore_patterns_have_changed == Some(false) |
|
308 | && self.ignore_patterns_have_changed == Some(false) | |
294 | { |
|
309 | { | |
295 | if let Some(cached_mtime) = cached_directory_mtime { |
|
310 | if let Some(cached_mtime) = cached_directory_mtime { | |
296 | // The dirstate contains a cached mtime for this directory, set |
|
311 | // The dirstate contains a cached mtime for this directory, set | |
297 | // by a previous run of the `status` algorithm which found this |
|
312 | // by a previous run of the `status` algorithm which found this | |
298 | // directory eligible for `read_dir` caching. |
|
313 | // directory eligible for `read_dir` caching. | |
299 | if let Some(meta) = directory_metadata { |
|
314 | if let Some(meta) = directory_metadata { | |
300 | if cached_mtime |
|
315 | if cached_mtime | |
301 | .likely_equal_to_mtime_of(meta) |
|
316 | .likely_equal_to_mtime_of(meta) | |
302 | .unwrap_or(false) |
|
317 | .unwrap_or(false) | |
303 | { |
|
318 | { | |
304 | // The mtime of that directory has not changed |
|
319 | // The mtime of that directory has not changed | |
305 | // since then, which means that the results of |
|
320 | // since then, which means that the results of | |
306 | // `read_dir` should also be unchanged. |
|
321 | // `read_dir` should also be unchanged. | |
307 | return true; |
|
322 | return true; | |
308 | } |
|
323 | } | |
309 | } |
|
324 | } | |
310 | } |
|
325 | } | |
311 | } |
|
326 | } | |
312 | false |
|
327 | false | |
313 | } |
|
328 | } | |
314 |
|
329 | |||
315 | /// Returns whether all child entries of the filesystem directory have a |
|
330 | /// Returns whether all child entries of the filesystem directory have a | |
316 | /// corresponding dirstate node or are ignored. |
|
331 | /// corresponding dirstate node or are ignored. | |
317 | fn traverse_fs_directory_and_dirstate( |
|
332 | fn traverse_fs_directory_and_dirstate( | |
318 | &self, |
|
333 | &self, | |
319 | has_ignored_ancestor: bool, |
|
334 | has_ignored_ancestor: bool, | |
320 | dirstate_nodes: ChildNodesRef<'tree, 'on_disk>, |
|
335 | dirstate_nodes: ChildNodesRef<'tree, 'on_disk>, | |
321 | directory_hg_path: &BorrowedPath<'tree, 'on_disk>, |
|
336 | directory_hg_path: &BorrowedPath<'tree, 'on_disk>, | |
322 | directory_fs_path: &Path, |
|
337 | directory_fs_path: &Path, | |
323 | directory_metadata: Option<&std::fs::Metadata>, |
|
338 | directory_metadata: Option<&std::fs::Metadata>, | |
324 | cached_directory_mtime: Option<TruncatedTimestamp>, |
|
339 | cached_directory_mtime: Option<TruncatedTimestamp>, | |
325 | is_at_repo_root: bool, |
|
340 | is_at_repo_root: bool, | |
326 | ) -> Result<bool, DirstateV2ParseError> { |
|
341 | ) -> Result<bool, DirstateV2ParseError> { | |
327 | if self.can_skip_fs_readdir(directory_metadata, cached_directory_mtime) |
|
342 | if self.can_skip_fs_readdir(directory_metadata, cached_directory_mtime) | |
328 | { |
|
343 | { | |
329 | dirstate_nodes |
|
344 | dirstate_nodes | |
330 | .par_iter() |
|
345 | .par_iter() | |
331 | .map(|dirstate_node| { |
|
346 | .map(|dirstate_node| { | |
332 | let fs_path = directory_fs_path.join(get_path_from_bytes( |
|
347 | let fs_path = directory_fs_path.join(get_path_from_bytes( | |
333 | dirstate_node.base_name(self.dmap.on_disk)?.as_bytes(), |
|
348 | dirstate_node.base_name(self.dmap.on_disk)?.as_bytes(), | |
334 | )); |
|
349 | )); | |
335 | match std::fs::symlink_metadata(&fs_path) { |
|
350 | match std::fs::symlink_metadata(&fs_path) { | |
336 | Ok(fs_metadata) => self.traverse_fs_and_dirstate( |
|
351 | Ok(fs_metadata) => self.traverse_fs_and_dirstate( | |
337 | &fs_path, |
|
352 | &fs_path, | |
338 | &fs_metadata, |
|
353 | &fs_metadata, | |
339 | dirstate_node, |
|
354 | dirstate_node, | |
340 | has_ignored_ancestor, |
|
355 | has_ignored_ancestor, | |
341 | ), |
|
356 | ), | |
342 | Err(e) if e.kind() == std::io::ErrorKind::NotFound => { |
|
357 | Err(e) if e.kind() == std::io::ErrorKind::NotFound => { | |
343 | self.traverse_dirstate_only(dirstate_node) |
|
358 | self.traverse_dirstate_only(dirstate_node) | |
344 | } |
|
359 | } | |
345 | Err(error) => { |
|
360 | Err(error) => { | |
346 | let hg_path = |
|
361 | let hg_path = | |
347 | dirstate_node.full_path(self.dmap.on_disk)?; |
|
362 | dirstate_node.full_path(self.dmap.on_disk)?; | |
348 | Ok(self.io_error(error, hg_path)) |
|
363 | Ok(self.io_error(error, hg_path)) | |
349 | } |
|
364 | } | |
350 | } |
|
365 | } | |
351 | }) |
|
366 | }) | |
352 | .collect::<Result<_, _>>()?; |
|
367 | .collect::<Result<_, _>>()?; | |
353 |
|
368 | |||
354 | // We don’t know, so conservatively say this isn’t the case |
|
369 | // We don’t know, so conservatively say this isn’t the case | |
355 | let children_all_have_dirstate_node_or_are_ignored = false; |
|
370 | let children_all_have_dirstate_node_or_are_ignored = false; | |
356 |
|
371 | |||
357 | return Ok(children_all_have_dirstate_node_or_are_ignored); |
|
372 | return Ok(children_all_have_dirstate_node_or_are_ignored); | |
358 | } |
|
373 | } | |
359 |
|
374 | |||
360 | let mut fs_entries = if let Ok(entries) = self.read_dir( |
|
375 | let mut fs_entries = if let Ok(entries) = self.read_dir( | |
361 | directory_hg_path, |
|
376 | directory_hg_path, | |
362 | directory_fs_path, |
|
377 | directory_fs_path, | |
363 | is_at_repo_root, |
|
378 | is_at_repo_root, | |
364 | ) { |
|
379 | ) { | |
365 | entries |
|
380 | entries | |
366 | } else { |
|
381 | } else { | |
367 | // Treat an unreadable directory (typically because of insufficient |
|
382 | // Treat an unreadable directory (typically because of insufficient | |
368 | // permissions) like an empty directory. `self.read_dir` has |
|
383 | // permissions) like an empty directory. `self.read_dir` has | |
369 | // already called `self.io_error` so a warning will be emitted. |
|
384 | // already called `self.io_error` so a warning will be emitted. | |
370 | Vec::new() |
|
385 | Vec::new() | |
371 | }; |
|
386 | }; | |
372 |
|
387 | |||
373 | // `merge_join_by` requires both its input iterators to be sorted: |
|
388 | // `merge_join_by` requires both its input iterators to be sorted: | |
374 |
|
389 | |||
375 | let dirstate_nodes = dirstate_nodes.sorted(); |
|
390 | let dirstate_nodes = dirstate_nodes.sorted(); | |
376 | // `sort_unstable_by_key` doesn’t allow keys borrowing from the value: |
|
391 | // `sort_unstable_by_key` doesn’t allow keys borrowing from the value: | |
377 | // https://github.com/rust-lang/rust/issues/34162 |
|
392 | // https://github.com/rust-lang/rust/issues/34162 | |
378 | fs_entries.sort_unstable_by(|e1, e2| e1.base_name.cmp(&e2.base_name)); |
|
393 | fs_entries.sort_unstable_by(|e1, e2| e1.base_name.cmp(&e2.base_name)); | |
379 |
|
394 | |||
380 | // Propagate here any error that would happen inside the comparison |
|
395 | // Propagate here any error that would happen inside the comparison | |
381 | // callback below |
|
396 | // callback below | |
382 | for dirstate_node in &dirstate_nodes { |
|
397 | for dirstate_node in &dirstate_nodes { | |
383 | dirstate_node.base_name(self.dmap.on_disk)?; |
|
398 | dirstate_node.base_name(self.dmap.on_disk)?; | |
384 | } |
|
399 | } | |
385 | itertools::merge_join_by( |
|
400 | itertools::merge_join_by( | |
386 | dirstate_nodes, |
|
401 | dirstate_nodes, | |
387 | &fs_entries, |
|
402 | &fs_entries, | |
388 | |dirstate_node, fs_entry| { |
|
403 | |dirstate_node, fs_entry| { | |
389 | // This `unwrap` never panics because we already propagated |
|
404 | // This `unwrap` never panics because we already propagated | |
390 | // those errors above |
|
405 | // those errors above | |
391 | dirstate_node |
|
406 | dirstate_node | |
392 | .base_name(self.dmap.on_disk) |
|
407 | .base_name(self.dmap.on_disk) | |
393 | .unwrap() |
|
408 | .unwrap() | |
394 | .cmp(&fs_entry.base_name) |
|
409 | .cmp(&fs_entry.base_name) | |
395 | }, |
|
410 | }, | |
396 | ) |
|
411 | ) | |
397 | .par_bridge() |
|
412 | .par_bridge() | |
398 | .map(|pair| { |
|
413 | .map(|pair| { | |
399 | use itertools::EitherOrBoth::*; |
|
414 | use itertools::EitherOrBoth::*; | |
400 | let has_dirstate_node_or_is_ignored; |
|
415 | let has_dirstate_node_or_is_ignored; | |
401 | match pair { |
|
416 | match pair { | |
402 | Both(dirstate_node, fs_entry) => { |
|
417 | Both(dirstate_node, fs_entry) => { | |
403 | self.traverse_fs_and_dirstate( |
|
418 | self.traverse_fs_and_dirstate( | |
404 | &fs_entry.full_path, |
|
419 | &fs_entry.full_path, | |
405 | &fs_entry.metadata, |
|
420 | &fs_entry.metadata, | |
406 | dirstate_node, |
|
421 | dirstate_node, | |
407 | has_ignored_ancestor, |
|
422 | has_ignored_ancestor, | |
408 | )?; |
|
423 | )?; | |
409 | has_dirstate_node_or_is_ignored = true |
|
424 | has_dirstate_node_or_is_ignored = true | |
410 | } |
|
425 | } | |
411 | Left(dirstate_node) => { |
|
426 | Left(dirstate_node) => { | |
412 | self.traverse_dirstate_only(dirstate_node)?; |
|
427 | self.traverse_dirstate_only(dirstate_node)?; | |
413 | has_dirstate_node_or_is_ignored = true; |
|
428 | has_dirstate_node_or_is_ignored = true; | |
414 | } |
|
429 | } | |
415 | Right(fs_entry) => { |
|
430 | Right(fs_entry) => { | |
416 | has_dirstate_node_or_is_ignored = self.traverse_fs_only( |
|
431 | has_dirstate_node_or_is_ignored = self.traverse_fs_only( | |
417 | has_ignored_ancestor, |
|
432 | has_ignored_ancestor, | |
418 | directory_hg_path, |
|
433 | directory_hg_path, | |
419 | fs_entry, |
|
434 | fs_entry, | |
420 | ) |
|
435 | ) | |
421 | } |
|
436 | } | |
422 | } |
|
437 | } | |
423 | Ok(has_dirstate_node_or_is_ignored) |
|
438 | Ok(has_dirstate_node_or_is_ignored) | |
424 | }) |
|
439 | }) | |
425 | .try_reduce(|| true, |a, b| Ok(a && b)) |
|
440 | .try_reduce(|| true, |a, b| Ok(a && b)) | |
426 | } |
|
441 | } | |
427 |
|
442 | |||
428 | fn traverse_fs_and_dirstate( |
|
443 | fn traverse_fs_and_dirstate( | |
429 | &self, |
|
444 | &self, | |
430 | fs_path: &Path, |
|
445 | fs_path: &Path, | |
431 | fs_metadata: &std::fs::Metadata, |
|
446 | fs_metadata: &std::fs::Metadata, | |
432 | dirstate_node: NodeRef<'tree, 'on_disk>, |
|
447 | dirstate_node: NodeRef<'tree, 'on_disk>, | |
433 | has_ignored_ancestor: bool, |
|
448 | has_ignored_ancestor: bool, | |
434 | ) -> Result<(), DirstateV2ParseError> { |
|
449 | ) -> Result<(), DirstateV2ParseError> { | |
435 | self.check_for_outdated_directory_cache(&dirstate_node)?; |
|
450 | self.check_for_outdated_directory_cache(&dirstate_node)?; | |
436 | let hg_path = &dirstate_node.full_path_borrowed(self.dmap.on_disk)?; |
|
451 | let hg_path = &dirstate_node.full_path_borrowed(self.dmap.on_disk)?; | |
437 | let file_type = fs_metadata.file_type(); |
|
452 | let file_type = fs_metadata.file_type(); | |
438 | let file_or_symlink = file_type.is_file() || file_type.is_symlink(); |
|
453 | let file_or_symlink = file_type.is_file() || file_type.is_symlink(); | |
439 | if !file_or_symlink { |
|
454 | if !file_or_symlink { | |
440 | // If we previously had a file here, it was removed (with |
|
455 | // If we previously had a file here, it was removed (with | |
441 | // `hg rm` or similar) or deleted before it could be |
|
456 | // `hg rm` or similar) or deleted before it could be | |
442 | // replaced by a directory or something else. |
|
457 | // replaced by a directory or something else. | |
443 | self.mark_removed_or_deleted_if_file(&dirstate_node)?; |
|
458 | self.mark_removed_or_deleted_if_file(&dirstate_node)?; | |
444 | } |
|
459 | } | |
445 | if file_type.is_dir() { |
|
460 | if file_type.is_dir() { | |
446 | if self.options.collect_traversed_dirs { |
|
461 | if self.options.collect_traversed_dirs { | |
447 | self.outcome |
|
462 | self.outcome | |
448 | .lock() |
|
463 | .lock() | |
449 | .unwrap() |
|
464 | .unwrap() | |
450 | .traversed |
|
465 | .traversed | |
451 | .push(hg_path.detach_from_tree()) |
|
466 | .push(hg_path.detach_from_tree()) | |
452 | } |
|
467 | } | |
453 | let is_ignored = has_ignored_ancestor || (self.ignore_fn)(hg_path); |
|
468 | let is_ignored = has_ignored_ancestor || (self.ignore_fn)(hg_path); | |
454 | let is_at_repo_root = false; |
|
469 | let is_at_repo_root = false; | |
455 | let children_all_have_dirstate_node_or_are_ignored = self |
|
470 | let children_all_have_dirstate_node_or_are_ignored = self | |
456 | .traverse_fs_directory_and_dirstate( |
|
471 | .traverse_fs_directory_and_dirstate( | |
457 | is_ignored, |
|
472 | is_ignored, | |
458 | dirstate_node.children(self.dmap.on_disk)?, |
|
473 | dirstate_node.children(self.dmap.on_disk)?, | |
459 | hg_path, |
|
474 | hg_path, | |
460 | fs_path, |
|
475 | fs_path, | |
461 | Some(fs_metadata), |
|
476 | Some(fs_metadata), | |
462 | dirstate_node.cached_directory_mtime()?, |
|
477 | dirstate_node.cached_directory_mtime()?, | |
463 | is_at_repo_root, |
|
478 | is_at_repo_root, | |
464 | )?; |
|
479 | )?; | |
465 | self.maybe_save_directory_mtime( |
|
480 | self.maybe_save_directory_mtime( | |
466 | children_all_have_dirstate_node_or_are_ignored, |
|
481 | children_all_have_dirstate_node_or_are_ignored, | |
467 | fs_metadata, |
|
482 | fs_metadata, | |
468 | dirstate_node, |
|
483 | dirstate_node, | |
469 | )? |
|
484 | )? | |
470 | } else { |
|
485 | } else { | |
471 | if file_or_symlink && self.matcher.matches(hg_path) { |
|
486 | if file_or_symlink && self.matcher.matches(hg_path) { | |
472 | if let Some(state) = dirstate_node.state()? { |
|
487 | if let Some(state) = dirstate_node.state()? { | |
473 | match state { |
|
488 | match state { | |
474 | EntryState::Added => { |
|
489 | EntryState::Added => { | |
475 | self.push_outcome(Outcome::Added, &dirstate_node)? |
|
490 | self.push_outcome(Outcome::Added, &dirstate_node)? | |
476 | } |
|
491 | } | |
477 | EntryState::Removed => self |
|
492 | EntryState::Removed => self | |
478 | .push_outcome(Outcome::Removed, &dirstate_node)?, |
|
493 | .push_outcome(Outcome::Removed, &dirstate_node)?, | |
479 | EntryState::Merged => self |
|
494 | EntryState::Merged => self | |
480 | .push_outcome(Outcome::Modified, &dirstate_node)?, |
|
495 | .push_outcome(Outcome::Modified, &dirstate_node)?, | |
481 | EntryState::Normal => self |
|
496 | EntryState::Normal => self | |
482 | .handle_normal_file(&dirstate_node, fs_metadata)?, |
|
497 | .handle_normal_file(&dirstate_node, fs_metadata)?, | |
483 | } |
|
498 | } | |
484 | } else { |
|
499 | } else { | |
485 | // `node.entry.is_none()` indicates a "directory" |
|
500 | // `node.entry.is_none()` indicates a "directory" | |
486 | // node, but the filesystem has a file |
|
501 | // node, but the filesystem has a file | |
487 | self.mark_unknown_or_ignored( |
|
502 | self.mark_unknown_or_ignored( | |
488 | has_ignored_ancestor, |
|
503 | has_ignored_ancestor, | |
489 | hg_path, |
|
504 | hg_path, | |
490 | ); |
|
505 | ); | |
491 | } |
|
506 | } | |
492 | } |
|
507 | } | |
493 |
|
508 | |||
494 | for child_node in dirstate_node.children(self.dmap.on_disk)?.iter() |
|
509 | for child_node in dirstate_node.children(self.dmap.on_disk)?.iter() | |
495 | { |
|
510 | { | |
496 | self.traverse_dirstate_only(child_node)? |
|
511 | self.traverse_dirstate_only(child_node)? | |
497 | } |
|
512 | } | |
498 | } |
|
513 | } | |
499 | Ok(()) |
|
514 | Ok(()) | |
500 | } |
|
515 | } | |
501 |
|
516 | |||
502 | fn maybe_save_directory_mtime( |
|
517 | fn maybe_save_directory_mtime( | |
503 | &self, |
|
518 | &self, | |
504 | children_all_have_dirstate_node_or_are_ignored: bool, |
|
519 | children_all_have_dirstate_node_or_are_ignored: bool, | |
505 | directory_metadata: &std::fs::Metadata, |
|
520 | directory_metadata: &std::fs::Metadata, | |
506 | dirstate_node: NodeRef<'tree, 'on_disk>, |
|
521 | dirstate_node: NodeRef<'tree, 'on_disk>, | |
507 | ) -> Result<(), DirstateV2ParseError> { |
|
522 | ) -> Result<(), DirstateV2ParseError> { | |
508 | if !children_all_have_dirstate_node_or_are_ignored { |
|
523 | if !children_all_have_dirstate_node_or_are_ignored { | |
509 | return Ok(()); |
|
524 | return Ok(()); | |
510 | } |
|
525 | } | |
511 | // All filesystem directory entries from `read_dir` have a |
|
526 | // All filesystem directory entries from `read_dir` have a | |
512 | // corresponding node in the dirstate, so we can reconstitute the |
|
527 | // corresponding node in the dirstate, so we can reconstitute the | |
513 | // names of those entries without calling `read_dir` again. |
|
528 | // names of those entries without calling `read_dir` again. | |
514 |
|
529 | |||
515 | // TODO: use let-else here and below when available: |
|
530 | // TODO: use let-else here and below when available: | |
516 | // https://github.com/rust-lang/rust/issues/87335 |
|
531 | // https://github.com/rust-lang/rust/issues/87335 | |
517 | let status_start = if let Some(status_start) = |
|
532 | let status_start = if let Some(status_start) = | |
518 | &self.filesystem_time_at_status_start |
|
533 | &self.filesystem_time_at_status_start | |
519 | { |
|
534 | { | |
520 | status_start |
|
535 | status_start | |
521 | } else { |
|
536 | } else { | |
522 | return Ok(()); |
|
537 | return Ok(()); | |
523 | }; |
|
538 | }; | |
524 |
|
539 | |||
525 | // Although the Rust standard library’s `SystemTime` type |
|
540 | // Although the Rust standard library’s `SystemTime` type | |
526 | // has nanosecond precision, the times reported for a |
|
541 | // has nanosecond precision, the times reported for a | |
527 | // directory’s (or file’s) modified time may have lower |
|
542 | // directory’s (or file’s) modified time may have lower | |
528 | // resolution based on the filesystem (for example ext3 |
|
543 | // resolution based on the filesystem (for example ext3 | |
529 | // only stores integer seconds), kernel (see |
|
544 | // only stores integer seconds), kernel (see | |
530 | // https://stackoverflow.com/a/14393315/1162888), etc. |
|
545 | // https://stackoverflow.com/a/14393315/1162888), etc. | |
531 | let directory_mtime = if let Ok(option) = |
|
546 | let directory_mtime = if let Ok(option) = | |
532 | TruncatedTimestamp::for_reliable_mtime_of( |
|
547 | TruncatedTimestamp::for_reliable_mtime_of( | |
533 | directory_metadata, |
|
548 | directory_metadata, | |
534 | status_start, |
|
549 | status_start, | |
535 | ) { |
|
550 | ) { | |
536 | if let Some(directory_mtime) = option { |
|
551 | if let Some(directory_mtime) = option { | |
537 | directory_mtime |
|
552 | directory_mtime | |
538 | } else { |
|
553 | } else { | |
539 | // The directory was modified too recently, |
|
554 | // The directory was modified too recently, | |
540 | // don’t cache its `read_dir` results. |
|
555 | // don’t cache its `read_dir` results. | |
541 | // |
|
556 | // | |
542 | // 1. A change to this directory (direct child was |
|
557 | // 1. A change to this directory (direct child was | |
543 | // added or removed) cause its mtime to be set |
|
558 | // added or removed) cause its mtime to be set | |
544 | // (possibly truncated) to `directory_mtime` |
|
559 | // (possibly truncated) to `directory_mtime` | |
545 | // 2. This `status` algorithm calls `read_dir` |
|
560 | // 2. This `status` algorithm calls `read_dir` | |
546 | // 3. An other change is made to the same directory is |
|
561 | // 3. An other change is made to the same directory is | |
547 | // made so that calling `read_dir` agin would give |
|
562 | // made so that calling `read_dir` agin would give | |
548 | // different results, but soon enough after 1. that |
|
563 | // different results, but soon enough after 1. that | |
549 | // the mtime stays the same |
|
564 | // the mtime stays the same | |
550 | // |
|
565 | // | |
551 | // On a system where the time resolution poor, this |
|
566 | // On a system where the time resolution poor, this | |
552 | // scenario is not unlikely if all three steps are caused |
|
567 | // scenario is not unlikely if all three steps are caused | |
553 | // by the same script. |
|
568 | // by the same script. | |
554 | return Ok(()); |
|
569 | return Ok(()); | |
555 | } |
|
570 | } | |
556 | } else { |
|
571 | } else { | |
557 | // OS/libc does not support mtime? |
|
572 | // OS/libc does not support mtime? | |
558 | return Ok(()); |
|
573 | return Ok(()); | |
559 | }; |
|
574 | }; | |
560 | // We’ve observed (through `status_start`) that time has |
|
575 | // We’ve observed (through `status_start`) that time has | |
561 | // “progressed” since `directory_mtime`, so any further |
|
576 | // “progressed” since `directory_mtime`, so any further | |
562 | // change to this directory is extremely likely to cause a |
|
577 | // change to this directory is extremely likely to cause a | |
563 | // different mtime. |
|
578 | // different mtime. | |
564 | // |
|
579 | // | |
565 | // Having the same mtime again is not entirely impossible |
|
580 | // Having the same mtime again is not entirely impossible | |
566 | // since the system clock is not monotonous. It could jump |
|
581 | // since the system clock is not monotonous. It could jump | |
567 | // backward to some point before `directory_mtime`, then a |
|
582 | // backward to some point before `directory_mtime`, then a | |
568 | // directory change could potentially happen during exactly |
|
583 | // directory change could potentially happen during exactly | |
569 | // the wrong tick. |
|
584 | // the wrong tick. | |
570 | // |
|
585 | // | |
571 | // We deem this scenario (unlike the previous one) to be |
|
586 | // We deem this scenario (unlike the previous one) to be | |
572 | // unlikely enough in practice. |
|
587 | // unlikely enough in practice. | |
573 |
|
588 | |||
574 | let is_up_to_date = |
|
589 | let is_up_to_date = | |
575 | if let Some(cached) = dirstate_node.cached_directory_mtime()? { |
|
590 | if let Some(cached) = dirstate_node.cached_directory_mtime()? { | |
576 | cached.likely_equal(directory_mtime) |
|
591 | cached.likely_equal(directory_mtime) | |
577 | } else { |
|
592 | } else { | |
578 | false |
|
593 | false | |
579 | }; |
|
594 | }; | |
580 | if !is_up_to_date { |
|
595 | if !is_up_to_date { | |
581 | let hg_path = dirstate_node |
|
596 | let hg_path = dirstate_node | |
582 | .full_path_borrowed(self.dmap.on_disk)? |
|
597 | .full_path_borrowed(self.dmap.on_disk)? | |
583 | .detach_from_tree(); |
|
598 | .detach_from_tree(); | |
584 | self.new_cachable_directories |
|
599 | self.new_cachable_directories | |
585 | .lock() |
|
600 | .lock() | |
586 | .unwrap() |
|
601 | .unwrap() | |
587 | .push((hg_path, directory_mtime)) |
|
602 | .push((hg_path, directory_mtime)) | |
588 | } |
|
603 | } | |
589 | Ok(()) |
|
604 | Ok(()) | |
590 | } |
|
605 | } | |
591 |
|
606 | |||
592 | /// A file with `EntryState::Normal` in the dirstate was found in the |
|
607 | /// A file with `EntryState::Normal` in the dirstate was found in the | |
593 | /// filesystem |
|
608 | /// filesystem | |
594 | fn handle_normal_file( |
|
609 | fn handle_normal_file( | |
595 | &self, |
|
610 | &self, | |
596 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
611 | dirstate_node: &NodeRef<'tree, 'on_disk>, | |
597 | fs_metadata: &std::fs::Metadata, |
|
612 | fs_metadata: &std::fs::Metadata, | |
598 | ) -> Result<(), DirstateV2ParseError> { |
|
613 | ) -> Result<(), DirstateV2ParseError> { | |
599 | // Keep the low 31 bits |
|
614 | // Keep the low 31 bits | |
600 | fn truncate_u64(value: u64) -> i32 { |
|
615 | fn truncate_u64(value: u64) -> i32 { | |
601 | (value & 0x7FFF_FFFF) as i32 |
|
616 | (value & 0x7FFF_FFFF) as i32 | |
602 | } |
|
617 | } | |
603 |
|
618 | |||
604 | let entry = dirstate_node |
|
619 | let entry = dirstate_node | |
605 | .entry()? |
|
620 | .entry()? | |
606 | .expect("handle_normal_file called with entry-less node"); |
|
621 | .expect("handle_normal_file called with entry-less node"); | |
607 | let mode_changed = |
|
622 | let mode_changed = | |
608 | || self.options.check_exec && entry.mode_changed(fs_metadata); |
|
623 | || self.options.check_exec && entry.mode_changed(fs_metadata); | |
609 | let size = entry.size(); |
|
624 | let size = entry.size(); | |
610 | let size_changed = size != truncate_u64(fs_metadata.len()); |
|
625 | let size_changed = size != truncate_u64(fs_metadata.len()); | |
611 | if size >= 0 && size_changed && fs_metadata.file_type().is_symlink() { |
|
626 | if size >= 0 && size_changed && fs_metadata.file_type().is_symlink() { | |
612 | // issue6456: Size returned may be longer due to encryption |
|
627 | // issue6456: Size returned may be longer due to encryption | |
613 | // on EXT-4 fscrypt. TODO maybe only do it on EXT4? |
|
628 | // on EXT-4 fscrypt. TODO maybe only do it on EXT4? | |
614 | self.push_outcome(Outcome::Unsure, dirstate_node)? |
|
629 | self.push_outcome(Outcome::Unsure, dirstate_node)? | |
615 | } else if dirstate_node.has_copy_source() |
|
630 | } else if dirstate_node.has_copy_source() | |
616 | || entry.is_from_other_parent() |
|
631 | || entry.is_from_other_parent() | |
617 | || (size >= 0 && (size_changed || mode_changed())) |
|
632 | || (size >= 0 && (size_changed || mode_changed())) | |
618 | { |
|
633 | { | |
619 | self.push_outcome(Outcome::Modified, dirstate_node)? |
|
634 | self.push_outcome(Outcome::Modified, dirstate_node)? | |
620 | } else { |
|
635 | } else { | |
621 | let mtime_looks_clean; |
|
636 | let mtime_looks_clean; | |
622 | if let Some(dirstate_mtime) = entry.truncated_mtime() { |
|
637 | if let Some(dirstate_mtime) = entry.truncated_mtime() { | |
623 | let fs_mtime = TruncatedTimestamp::for_mtime_of(fs_metadata) |
|
638 | let fs_mtime = TruncatedTimestamp::for_mtime_of(fs_metadata) | |
624 | .expect("OS/libc does not support mtime?"); |
|
639 | .expect("OS/libc does not support mtime?"); | |
625 | // There might be a change in the future if for example the |
|
640 | // There might be a change in the future if for example the | |
626 | // internal clock become off while process run, but this is a |
|
641 | // internal clock become off while process run, but this is a | |
627 | // case where the issues the user would face |
|
642 | // case where the issues the user would face | |
628 | // would be a lot worse and there is nothing we |
|
643 | // would be a lot worse and there is nothing we | |
629 | // can really do. |
|
644 | // can really do. | |
630 | mtime_looks_clean = fs_mtime.likely_equal(dirstate_mtime) |
|
645 | mtime_looks_clean = fs_mtime.likely_equal(dirstate_mtime) | |
631 | } else { |
|
646 | } else { | |
632 | // No mtime in the dirstate entry |
|
647 | // No mtime in the dirstate entry | |
633 | mtime_looks_clean = false |
|
648 | mtime_looks_clean = false | |
634 | }; |
|
649 | }; | |
635 | if !mtime_looks_clean { |
|
650 | if !mtime_looks_clean { | |
636 | self.push_outcome(Outcome::Unsure, dirstate_node)? |
|
651 | self.push_outcome(Outcome::Unsure, dirstate_node)? | |
637 | } else if self.options.list_clean { |
|
652 | } else if self.options.list_clean { | |
638 | self.push_outcome(Outcome::Clean, dirstate_node)? |
|
653 | self.push_outcome(Outcome::Clean, dirstate_node)? | |
639 | } |
|
654 | } | |
640 | } |
|
655 | } | |
641 | Ok(()) |
|
656 | Ok(()) | |
642 | } |
|
657 | } | |
643 |
|
658 | |||
644 | /// A node in the dirstate tree has no corresponding filesystem entry |
|
659 | /// A node in the dirstate tree has no corresponding filesystem entry | |
645 | fn traverse_dirstate_only( |
|
660 | fn traverse_dirstate_only( | |
646 | &self, |
|
661 | &self, | |
647 | dirstate_node: NodeRef<'tree, 'on_disk>, |
|
662 | dirstate_node: NodeRef<'tree, 'on_disk>, | |
648 | ) -> Result<(), DirstateV2ParseError> { |
|
663 | ) -> Result<(), DirstateV2ParseError> { | |
649 | self.check_for_outdated_directory_cache(&dirstate_node)?; |
|
664 | self.check_for_outdated_directory_cache(&dirstate_node)?; | |
650 | self.mark_removed_or_deleted_if_file(&dirstate_node)?; |
|
665 | self.mark_removed_or_deleted_if_file(&dirstate_node)?; | |
651 | dirstate_node |
|
666 | dirstate_node | |
652 | .children(self.dmap.on_disk)? |
|
667 | .children(self.dmap.on_disk)? | |
653 | .par_iter() |
|
668 | .par_iter() | |
654 | .map(|child_node| self.traverse_dirstate_only(child_node)) |
|
669 | .map(|child_node| self.traverse_dirstate_only(child_node)) | |
655 | .collect() |
|
670 | .collect() | |
656 | } |
|
671 | } | |
657 |
|
672 | |||
658 | /// A node in the dirstate tree has no corresponding *file* on the |
|
673 | /// A node in the dirstate tree has no corresponding *file* on the | |
659 | /// filesystem |
|
674 | /// filesystem | |
660 | /// |
|
675 | /// | |
661 | /// Does nothing on a "directory" node |
|
676 | /// Does nothing on a "directory" node | |
662 | fn mark_removed_or_deleted_if_file( |
|
677 | fn mark_removed_or_deleted_if_file( | |
663 | &self, |
|
678 | &self, | |
664 | dirstate_node: &NodeRef<'tree, 'on_disk>, |
|
679 | dirstate_node: &NodeRef<'tree, 'on_disk>, | |
665 | ) -> Result<(), DirstateV2ParseError> { |
|
680 | ) -> Result<(), DirstateV2ParseError> { | |
666 | if let Some(state) = dirstate_node.state()? { |
|
681 | if let Some(state) = dirstate_node.state()? { | |
667 | let path = dirstate_node.full_path(self.dmap.on_disk)?; |
|
682 | let path = dirstate_node.full_path(self.dmap.on_disk)?; | |
668 | if self.matcher.matches(path) { |
|
683 | if self.matcher.matches(path) { | |
669 | if let EntryState::Removed = state { |
|
684 | if let EntryState::Removed = state { | |
670 | self.push_outcome(Outcome::Removed, dirstate_node)? |
|
685 | self.push_outcome(Outcome::Removed, dirstate_node)? | |
671 | } else { |
|
686 | } else { | |
672 | self.push_outcome(Outcome::Deleted, &dirstate_node)? |
|
687 | self.push_outcome(Outcome::Deleted, &dirstate_node)? | |
673 | } |
|
688 | } | |
674 | } |
|
689 | } | |
675 | } |
|
690 | } | |
676 | Ok(()) |
|
691 | Ok(()) | |
677 | } |
|
692 | } | |
678 |
|
693 | |||
679 | /// Something in the filesystem has no corresponding dirstate node |
|
694 | /// Something in the filesystem has no corresponding dirstate node | |
680 | /// |
|
695 | /// | |
681 | /// Returns whether that path is ignored |
|
696 | /// Returns whether that path is ignored | |
682 | fn traverse_fs_only( |
|
697 | fn traverse_fs_only( | |
683 | &self, |
|
698 | &self, | |
684 | has_ignored_ancestor: bool, |
|
699 | has_ignored_ancestor: bool, | |
685 | directory_hg_path: &HgPath, |
|
700 | directory_hg_path: &HgPath, | |
686 | fs_entry: &DirEntry, |
|
701 | fs_entry: &DirEntry, | |
687 | ) -> bool { |
|
702 | ) -> bool { | |
688 | let hg_path = directory_hg_path.join(&fs_entry.base_name); |
|
703 | let hg_path = directory_hg_path.join(&fs_entry.base_name); | |
689 | let file_type = fs_entry.metadata.file_type(); |
|
704 | let file_type = fs_entry.metadata.file_type(); | |
690 | let file_or_symlink = file_type.is_file() || file_type.is_symlink(); |
|
705 | let file_or_symlink = file_type.is_file() || file_type.is_symlink(); | |
691 | if file_type.is_dir() { |
|
706 | if file_type.is_dir() { | |
692 | let is_ignored = |
|
707 | let is_ignored = | |
693 | has_ignored_ancestor || (self.ignore_fn)(&hg_path); |
|
708 | has_ignored_ancestor || (self.ignore_fn)(&hg_path); | |
694 | let traverse_children = if is_ignored { |
|
709 | let traverse_children = if is_ignored { | |
695 | // Descendants of an ignored directory are all ignored |
|
710 | // Descendants of an ignored directory are all ignored | |
696 | self.options.list_ignored |
|
711 | self.options.list_ignored | |
697 | } else { |
|
712 | } else { | |
698 | // Descendants of an unknown directory may be either unknown or |
|
713 | // Descendants of an unknown directory may be either unknown or | |
699 | // ignored |
|
714 | // ignored | |
700 | self.options.list_unknown || self.options.list_ignored |
|
715 | self.options.list_unknown || self.options.list_ignored | |
701 | }; |
|
716 | }; | |
702 | if traverse_children { |
|
717 | if traverse_children { | |
703 | let is_at_repo_root = false; |
|
718 | let is_at_repo_root = false; | |
704 | if let Ok(children_fs_entries) = self.read_dir( |
|
719 | if let Ok(children_fs_entries) = self.read_dir( | |
705 | &hg_path, |
|
720 | &hg_path, | |
706 | &fs_entry.full_path, |
|
721 | &fs_entry.full_path, | |
707 | is_at_repo_root, |
|
722 | is_at_repo_root, | |
708 | ) { |
|
723 | ) { | |
709 | children_fs_entries.par_iter().for_each(|child_fs_entry| { |
|
724 | children_fs_entries.par_iter().for_each(|child_fs_entry| { | |
710 | self.traverse_fs_only( |
|
725 | self.traverse_fs_only( | |
711 | is_ignored, |
|
726 | is_ignored, | |
712 | &hg_path, |
|
727 | &hg_path, | |
713 | child_fs_entry, |
|
728 | child_fs_entry, | |
714 | ); |
|
729 | ); | |
715 | }) |
|
730 | }) | |
716 | } |
|
731 | } | |
717 | } |
|
732 | } | |
718 | if self.options.collect_traversed_dirs { |
|
733 | if self.options.collect_traversed_dirs { | |
719 | self.outcome.lock().unwrap().traversed.push(hg_path.into()) |
|
734 | self.outcome.lock().unwrap().traversed.push(hg_path.into()) | |
720 | } |
|
735 | } | |
721 | is_ignored |
|
736 | is_ignored | |
722 | } else { |
|
737 | } else { | |
723 | if file_or_symlink { |
|
738 | if file_or_symlink { | |
724 | if self.matcher.matches(&hg_path) { |
|
739 | if self.matcher.matches(&hg_path) { | |
725 | self.mark_unknown_or_ignored( |
|
740 | self.mark_unknown_or_ignored( | |
726 | has_ignored_ancestor, |
|
741 | has_ignored_ancestor, | |
727 | &BorrowedPath::InMemory(&hg_path), |
|
742 | &BorrowedPath::InMemory(&hg_path), | |
728 | ) |
|
743 | ) | |
729 | } else { |
|
744 | } else { | |
730 | // We haven’t computed whether this path is ignored. It |
|
745 | // We haven’t computed whether this path is ignored. It | |
731 | // might not be, and a future run of status might have a |
|
746 | // might not be, and a future run of status might have a | |
732 | // different matcher that matches it. So treat it as not |
|
747 | // different matcher that matches it. So treat it as not | |
733 | // ignored. That is, inhibit readdir caching of the parent |
|
748 | // ignored. That is, inhibit readdir caching of the parent | |
734 | // directory. |
|
749 | // directory. | |
735 | false |
|
750 | false | |
736 | } |
|
751 | } | |
737 | } else { |
|
752 | } else { | |
738 | // This is neither a directory, a plain file, or a symlink. |
|
753 | // This is neither a directory, a plain file, or a symlink. | |
739 | // Treat it like an ignored file. |
|
754 | // Treat it like an ignored file. | |
740 | true |
|
755 | true | |
741 | } |
|
756 | } | |
742 | } |
|
757 | } | |
743 | } |
|
758 | } | |
744 |
|
759 | |||
745 | /// Returns whether that path is ignored |
|
760 | /// Returns whether that path is ignored | |
746 | fn mark_unknown_or_ignored( |
|
761 | fn mark_unknown_or_ignored( | |
747 | &self, |
|
762 | &self, | |
748 | has_ignored_ancestor: bool, |
|
763 | has_ignored_ancestor: bool, | |
749 | hg_path: &BorrowedPath<'_, 'on_disk>, |
|
764 | hg_path: &BorrowedPath<'_, 'on_disk>, | |
750 | ) -> bool { |
|
765 | ) -> bool { | |
751 | let is_ignored = has_ignored_ancestor || (self.ignore_fn)(&hg_path); |
|
766 | let is_ignored = has_ignored_ancestor || (self.ignore_fn)(&hg_path); | |
752 | if is_ignored { |
|
767 | if is_ignored { | |
753 | if self.options.list_ignored { |
|
768 | if self.options.list_ignored { | |
754 | self.push_outcome_without_copy_source( |
|
769 | self.push_outcome_without_copy_source( | |
755 | Outcome::Ignored, |
|
770 | Outcome::Ignored, | |
756 | hg_path, |
|
771 | hg_path, | |
757 | ) |
|
772 | ) | |
758 | } |
|
773 | } | |
759 | } else { |
|
774 | } else { | |
760 | if self.options.list_unknown { |
|
775 | if self.options.list_unknown { | |
761 | self.push_outcome_without_copy_source( |
|
776 | self.push_outcome_without_copy_source( | |
762 | Outcome::Unknown, |
|
777 | Outcome::Unknown, | |
763 | hg_path, |
|
778 | hg_path, | |
764 | ) |
|
779 | ) | |
765 | } |
|
780 | } | |
766 | } |
|
781 | } | |
767 | is_ignored |
|
782 | is_ignored | |
768 | } |
|
783 | } | |
769 | } |
|
784 | } | |
770 |
|
785 | |||
771 | struct DirEntry { |
|
786 | struct DirEntry { | |
772 | base_name: HgPathBuf, |
|
787 | base_name: HgPathBuf, | |
773 | full_path: PathBuf, |
|
788 | full_path: PathBuf, | |
774 | metadata: std::fs::Metadata, |
|
789 | metadata: std::fs::Metadata, | |
775 | } |
|
790 | } | |
776 |
|
791 | |||
777 | impl DirEntry { |
|
792 | impl DirEntry { | |
778 | /// Returns **unsorted** entries in the given directory, with name and |
|
793 | /// Returns **unsorted** entries in the given directory, with name and | |
779 | /// metadata. |
|
794 | /// metadata. | |
780 | /// |
|
795 | /// | |
781 | /// If a `.hg` sub-directory is encountered: |
|
796 | /// If a `.hg` sub-directory is encountered: | |
782 | /// |
|
797 | /// | |
783 | /// * At the repository root, ignore that sub-directory |
|
798 | /// * At the repository root, ignore that sub-directory | |
784 | /// * Elsewhere, we’re listing the content of a sub-repo. Return an empty |
|
799 | /// * Elsewhere, we’re listing the content of a sub-repo. Return an empty | |
785 | /// list instead. |
|
800 | /// list instead. | |
786 | fn read_dir(path: &Path, is_at_repo_root: bool) -> io::Result<Vec<Self>> { |
|
801 | fn read_dir(path: &Path, is_at_repo_root: bool) -> io::Result<Vec<Self>> { | |
787 | // `read_dir` returns a "not found" error for the empty path |
|
802 | // `read_dir` returns a "not found" error for the empty path | |
788 | let at_cwd = path == Path::new(""); |
|
803 | let at_cwd = path == Path::new(""); | |
789 | let read_dir_path = if at_cwd { Path::new(".") } else { path }; |
|
804 | let read_dir_path = if at_cwd { Path::new(".") } else { path }; | |
790 | let mut results = Vec::new(); |
|
805 | let mut results = Vec::new(); | |
791 | for entry in read_dir_path.read_dir()? { |
|
806 | for entry in read_dir_path.read_dir()? { | |
792 | let entry = entry?; |
|
807 | let entry = entry?; | |
793 | let metadata = match entry.metadata() { |
|
808 | let metadata = match entry.metadata() { | |
794 | Ok(v) => v, |
|
809 | Ok(v) => v, | |
795 | Err(e) => { |
|
810 | Err(e) => { | |
796 | // race with file deletion? |
|
811 | // race with file deletion? | |
797 | if e.kind() == std::io::ErrorKind::NotFound { |
|
812 | if e.kind() == std::io::ErrorKind::NotFound { | |
798 | continue; |
|
813 | continue; | |
799 | } else { |
|
814 | } else { | |
800 | return Err(e); |
|
815 | return Err(e); | |
801 | } |
|
816 | } | |
802 | } |
|
817 | } | |
803 | }; |
|
818 | }; | |
804 | let file_name = entry.file_name(); |
|
819 | let file_name = entry.file_name(); | |
805 | // FIXME don't do this when cached |
|
820 | // FIXME don't do this when cached | |
806 | if file_name == ".hg" { |
|
821 | if file_name == ".hg" { | |
807 | if is_at_repo_root { |
|
822 | if is_at_repo_root { | |
808 | // Skip the repo’s own .hg (might be a symlink) |
|
823 | // Skip the repo’s own .hg (might be a symlink) | |
809 | continue; |
|
824 | continue; | |
810 | } else if metadata.is_dir() { |
|
825 | } else if metadata.is_dir() { | |
811 | // A .hg sub-directory at another location means a subrepo, |
|
826 | // A .hg sub-directory at another location means a subrepo, | |
812 | // skip it entirely. |
|
827 | // skip it entirely. | |
813 | return Ok(Vec::new()); |
|
828 | return Ok(Vec::new()); | |
814 | } |
|
829 | } | |
815 | } |
|
830 | } | |
816 | let full_path = if at_cwd { |
|
831 | let full_path = if at_cwd { | |
817 | file_name.clone().into() |
|
832 | file_name.clone().into() | |
818 | } else { |
|
833 | } else { | |
819 | entry.path() |
|
834 | entry.path() | |
820 | }; |
|
835 | }; | |
821 | let base_name = get_bytes_from_os_string(file_name).into(); |
|
836 | let base_name = get_bytes_from_os_string(file_name).into(); | |
822 | results.push(DirEntry { |
|
837 | results.push(DirEntry { | |
823 | base_name, |
|
838 | base_name, | |
824 | full_path, |
|
839 | full_path, | |
825 | metadata, |
|
840 | metadata, | |
826 | }) |
|
841 | }) | |
827 | } |
|
842 | } | |
828 | Ok(results) |
|
843 | Ok(results) | |
829 | } |
|
844 | } | |
830 | } |
|
845 | } | |
831 |
|
846 | |||
832 | /// Return the `mtime` of a temporary file newly-created in the `.hg` directory |
|
847 | /// Return the `mtime` of a temporary file newly-created in the `.hg` directory | |
833 | /// of the give repository. |
|
848 | /// of the give repository. | |
834 | /// |
|
849 | /// | |
835 | /// This is similar to `SystemTime::now()`, with the result truncated to the |
|
850 | /// This is similar to `SystemTime::now()`, with the result truncated to the | |
836 | /// same time resolution as other files’ modification times. Using `.hg` |
|
851 | /// same time resolution as other files’ modification times. Using `.hg` | |
837 | /// instead of the system’s default temporary directory (such as `/tmp`) makes |
|
852 | /// instead of the system’s default temporary directory (such as `/tmp`) makes | |
838 | /// it more likely the temporary file is in the same disk partition as contents |
|
853 | /// it more likely the temporary file is in the same disk partition as contents | |
839 | /// of the working directory, which can matter since different filesystems may |
|
854 | /// of the working directory, which can matter since different filesystems may | |
840 | /// store timestamps with different resolutions. |
|
855 | /// store timestamps with different resolutions. | |
841 | /// |
|
856 | /// | |
842 | /// This may fail, typically if we lack write permissions. In that case we |
|
857 | /// This may fail, typically if we lack write permissions. In that case we | |
843 | /// should continue the `status()` algoritm anyway and consider the current |
|
858 | /// should continue the `status()` algoritm anyway and consider the current | |
844 | /// date/time to be unknown. |
|
859 | /// date/time to be unknown. | |
845 | fn filesystem_now(repo_root: &Path) -> Result<SystemTime, io::Error> { |
|
860 | fn filesystem_now(repo_root: &Path) -> Result<SystemTime, io::Error> { | |
846 | tempfile::tempfile_in(repo_root.join(".hg"))? |
|
861 | tempfile::tempfile_in(repo_root.join(".hg"))? | |
847 | .metadata()? |
|
862 | .metadata()? | |
848 | .modified() |
|
863 | .modified() | |
849 | } |
|
864 | } |
General Comments 0
You need to be logged in to leave comments.
Login now