Show More
@@ -1,1069 +1,1069 b'' | |||||
1 | // Copyright 2018-2020 Georges Racinet <georges.racinet@octobus.net> |
|
1 | // Copyright 2018-2020 Georges Racinet <georges.racinet@octobus.net> | |
2 | // and Mercurial contributors |
|
2 | // and Mercurial contributors | |
3 | // |
|
3 | // | |
4 | // This software may be used and distributed according to the terms of the |
|
4 | // This software may be used and distributed according to the terms of the | |
5 | // GNU General Public License version 2 or any later version. |
|
5 | // GNU General Public License version 2 or any later version. | |
6 | //! Indexing facilities for fast retrieval of `Revision` from `Node` |
|
6 | //! Indexing facilities for fast retrieval of `Revision` from `Node` | |
7 | //! |
|
7 | //! | |
8 | //! This provides a variation on the 16-ary radix tree that is |
|
8 | //! This provides a variation on the 16-ary radix tree that is | |
9 | //! provided as "nodetree" in revlog.c, ready for append-only persistence |
|
9 | //! provided as "nodetree" in revlog.c, ready for append-only persistence | |
10 | //! on disk. |
|
10 | //! on disk. | |
11 | //! |
|
11 | //! | |
12 | //! Following existing implicit conventions, the "nodemap" terminology |
|
12 | //! Following existing implicit conventions, the "nodemap" terminology | |
13 | //! is used in a more abstract context. |
|
13 | //! is used in a more abstract context. | |
14 |
|
14 | |||
15 | use super::{ |
|
15 | use super::{ | |
16 | node::NULL_NODE, Node, NodePrefix, Revision, RevlogIndex, NULL_REVISION, |
|
16 | node::NULL_NODE, Node, NodePrefix, Revision, RevlogIndex, NULL_REVISION, | |
17 | }; |
|
17 | }; | |
18 |
|
18 | |||
19 | use bytes_cast::{unaligned, BytesCast}; |
|
19 | use bytes_cast::{unaligned, BytesCast}; | |
20 | use std::cmp::max; |
|
20 | use std::cmp::max; | |
21 | use std::fmt; |
|
21 | use std::fmt; | |
22 | use std::mem::{self, align_of, size_of}; |
|
22 | use std::mem::{self, align_of, size_of}; | |
23 | use std::ops::Deref; |
|
23 | use std::ops::Deref; | |
24 | use std::ops::Index; |
|
24 | use std::ops::Index; | |
25 |
|
25 | |||
26 | #[derive(Debug, PartialEq)] |
|
26 | #[derive(Debug, PartialEq)] | |
27 | pub enum NodeMapError { |
|
27 | pub enum NodeMapError { | |
28 | MultipleResults, |
|
28 | MultipleResults, | |
29 | /// A `Revision` stored in the nodemap could not be found in the index |
|
29 | /// A `Revision` stored in the nodemap could not be found in the index | |
30 | RevisionNotInIndex(Revision), |
|
30 | RevisionNotInIndex(Revision), | |
31 | } |
|
31 | } | |
32 |
|
32 | |||
33 | /// Mapping system from Mercurial nodes to revision numbers. |
|
33 | /// Mapping system from Mercurial nodes to revision numbers. | |
34 | /// |
|
34 | /// | |
35 | /// ## `RevlogIndex` and `NodeMap` |
|
35 | /// ## `RevlogIndex` and `NodeMap` | |
36 | /// |
|
36 | /// | |
37 | /// One way to think about their relationship is that |
|
37 | /// One way to think about their relationship is that | |
38 | /// the `NodeMap` is a prefix-oriented reverse index of the `Node` information |
|
38 | /// the `NodeMap` is a prefix-oriented reverse index of the `Node` information | |
39 | /// carried by a [`RevlogIndex`]. |
|
39 | /// carried by a [`RevlogIndex`]. | |
40 | /// |
|
40 | /// | |
41 | /// Many of the methods in this trait take a `RevlogIndex` argument |
|
41 | /// Many of the methods in this trait take a `RevlogIndex` argument | |
42 | /// which is used for validation of their results. This index must naturally |
|
42 | /// which is used for validation of their results. This index must naturally | |
43 | /// be the one the `NodeMap` is about, and it must be consistent. |
|
43 | /// be the one the `NodeMap` is about, and it must be consistent. | |
44 | /// |
|
44 | /// | |
45 | /// Notably, the `NodeMap` must not store |
|
45 | /// Notably, the `NodeMap` must not store | |
46 | /// information about more `Revision` values than there are in the index. |
|
46 | /// information about more `Revision` values than there are in the index. | |
47 | /// In these methods, an encountered `Revision` is not in the index, a |
|
47 | /// In these methods, an encountered `Revision` is not in the index, a | |
48 | /// [`RevisionNotInIndex`] error is returned. |
|
48 | /// [`RevisionNotInIndex`] error is returned. | |
49 | /// |
|
49 | /// | |
50 | /// In insert operations, the rule is thus that the `NodeMap` must always |
|
50 | /// In insert operations, the rule is thus that the `NodeMap` must always | |
51 | /// be updated after the `RevlogIndex` |
|
51 | /// be updated after the `RevlogIndex` | |
52 | /// be updated first, and the `NodeMap` second. |
|
52 | /// be updated first, and the `NodeMap` second. | |
53 | /// |
|
53 | /// | |
54 | /// [`RevisionNotInIndex`]: enum.NodeMapError.html#variant.RevisionNotInIndex |
|
54 | /// [`RevisionNotInIndex`]: enum.NodeMapError.html#variant.RevisionNotInIndex | |
55 | /// [`RevlogIndex`]: ../trait.RevlogIndex.html |
|
55 | /// [`RevlogIndex`]: ../trait.RevlogIndex.html | |
56 | pub trait NodeMap { |
|
56 | pub trait NodeMap { | |
57 | /// Find the unique `Revision` having the given `Node` |
|
57 | /// Find the unique `Revision` having the given `Node` | |
58 | /// |
|
58 | /// | |
59 | /// If no Revision matches the given `Node`, `Ok(None)` is returned. |
|
59 | /// If no Revision matches the given `Node`, `Ok(None)` is returned. | |
60 | fn find_node( |
|
60 | fn find_node( | |
61 | &self, |
|
61 | &self, | |
62 | index: &impl RevlogIndex, |
|
62 | index: &impl RevlogIndex, | |
63 | node: &Node, |
|
63 | node: &Node, | |
64 | ) -> Result<Option<Revision>, NodeMapError> { |
|
64 | ) -> Result<Option<Revision>, NodeMapError> { | |
65 | self.find_bin(index, node.into()) |
|
65 | self.find_bin(index, node.into()) | |
66 | } |
|
66 | } | |
67 |
|
67 | |||
68 | /// Find the unique Revision whose `Node` starts with a given binary prefix |
|
68 | /// Find the unique Revision whose `Node` starts with a given binary prefix | |
69 | /// |
|
69 | /// | |
70 | /// If no Revision matches the given prefix, `Ok(None)` is returned. |
|
70 | /// If no Revision matches the given prefix, `Ok(None)` is returned. | |
71 | /// |
|
71 | /// | |
72 | /// If several Revisions match the given prefix, a [`MultipleResults`] |
|
72 | /// If several Revisions match the given prefix, a [`MultipleResults`] | |
73 | /// error is returned. |
|
73 | /// error is returned. | |
74 | fn find_bin<'a>( |
|
74 | fn find_bin<'a>( | |
75 | &self, |
|
75 | &self, | |
76 | idx: &impl RevlogIndex, |
|
76 | idx: &impl RevlogIndex, | |
77 | prefix: NodePrefix, |
|
77 | prefix: NodePrefix, | |
78 | ) -> Result<Option<Revision>, NodeMapError>; |
|
78 | ) -> Result<Option<Revision>, NodeMapError>; | |
79 |
|
79 | |||
80 | /// Give the size of the shortest node prefix that determines |
|
80 | /// Give the size of the shortest node prefix that determines | |
81 | /// the revision uniquely. |
|
81 | /// the revision uniquely. | |
82 | /// |
|
82 | /// | |
83 | /// From a binary node prefix, if it is matched in the node map, this |
|
83 | /// From a binary node prefix, if it is matched in the node map, this | |
84 | /// returns the number of hexadecimal digits that would had sufficed |
|
84 | /// returns the number of hexadecimal digits that would had sufficed | |
85 | /// to find the revision uniquely. |
|
85 | /// to find the revision uniquely. | |
86 | /// |
|
86 | /// | |
87 | /// Returns `None` if no `Revision` could be found for the prefix. |
|
87 | /// Returns `None` if no `Revision` could be found for the prefix. | |
88 | /// |
|
88 | /// | |
89 | /// If several Revisions match the given prefix, a [`MultipleResults`] |
|
89 | /// If several Revisions match the given prefix, a [`MultipleResults`] | |
90 | /// error is returned. |
|
90 | /// error is returned. | |
91 | fn unique_prefix_len_bin<'a>( |
|
91 | fn unique_prefix_len_bin<'a>( | |
92 | &self, |
|
92 | &self, | |
93 | idx: &impl RevlogIndex, |
|
93 | idx: &impl RevlogIndex, | |
94 | node_prefix: NodePrefix, |
|
94 | node_prefix: NodePrefix, | |
95 | ) -> Result<Option<usize>, NodeMapError>; |
|
95 | ) -> Result<Option<usize>, NodeMapError>; | |
96 |
|
96 | |||
97 | /// Same as `unique_prefix_len_bin`, with a full `Node` as input |
|
97 | /// Same as `unique_prefix_len_bin`, with a full `Node` as input | |
98 | fn unique_prefix_len_node( |
|
98 | fn unique_prefix_len_node( | |
99 | &self, |
|
99 | &self, | |
100 | idx: &impl RevlogIndex, |
|
100 | idx: &impl RevlogIndex, | |
101 | node: &Node, |
|
101 | node: &Node, | |
102 | ) -> Result<Option<usize>, NodeMapError> { |
|
102 | ) -> Result<Option<usize>, NodeMapError> { | |
103 | self.unique_prefix_len_bin(idx, node.into()) |
|
103 | self.unique_prefix_len_bin(idx, node.into()) | |
104 | } |
|
104 | } | |
105 | } |
|
105 | } | |
106 |
|
106 | |||
107 | pub trait MutableNodeMap: NodeMap { |
|
107 | pub trait MutableNodeMap: NodeMap { | |
108 | fn insert<I: RevlogIndex>( |
|
108 | fn insert<I: RevlogIndex>( | |
109 | &mut self, |
|
109 | &mut self, | |
110 | index: &I, |
|
110 | index: &I, | |
111 | node: &Node, |
|
111 | node: &Node, | |
112 | rev: Revision, |
|
112 | rev: Revision, | |
113 | ) -> Result<(), NodeMapError>; |
|
113 | ) -> Result<(), NodeMapError>; | |
114 | } |
|
114 | } | |
115 |
|
115 | |||
116 | /// Low level NodeTree [`Blocks`] elements |
|
116 | /// Low level NodeTree [`Blocks`] elements | |
117 | /// |
|
117 | /// | |
118 | /// These are exactly as for instance on persistent storage. |
|
118 | /// These are exactly as for instance on persistent storage. | |
119 | type RawElement = unaligned::I32Be; |
|
119 | type RawElement = unaligned::I32Be; | |
120 |
|
120 | |||
121 | /// High level representation of values in NodeTree |
|
121 | /// High level representation of values in NodeTree | |
122 | /// [`Blocks`](struct.Block.html) |
|
122 | /// [`Blocks`](struct.Block.html) | |
123 | /// |
|
123 | /// | |
124 | /// This is the high level representation that most algorithms should |
|
124 | /// This is the high level representation that most algorithms should | |
125 | /// use. |
|
125 | /// use. | |
126 | #[derive(Clone, Debug, Eq, PartialEq)] |
|
126 | #[derive(Clone, Debug, Eq, PartialEq)] | |
127 | enum Element { |
|
127 | enum Element { | |
128 | Rev(Revision), |
|
128 | Rev(Revision), | |
129 | Block(usize), |
|
129 | Block(usize), | |
130 | None, |
|
130 | None, | |
131 | } |
|
131 | } | |
132 |
|
132 | |||
133 | impl From<RawElement> for Element { |
|
133 | impl From<RawElement> for Element { | |
134 | /// Conversion from low level representation, after endianness conversion. |
|
134 | /// Conversion from low level representation, after endianness conversion. | |
135 | /// |
|
135 | /// | |
136 | /// See [`Block`](struct.Block.html) for explanation about the encoding. |
|
136 | /// See [`Block`](struct.Block.html) for explanation about the encoding. | |
137 | fn from(raw: RawElement) -> Element { |
|
137 | fn from(raw: RawElement) -> Element { | |
138 | let int = raw.get(); |
|
138 | let int = raw.get(); | |
139 | if int >= 0 { |
|
139 | if int >= 0 { | |
140 | Element::Block(int as usize) |
|
140 | Element::Block(int as usize) | |
141 | } else if int == -1 { |
|
141 | } else if int == -1 { | |
142 | Element::None |
|
142 | Element::None | |
143 | } else { |
|
143 | } else { | |
144 | Element::Rev(-int - 2) |
|
144 | Element::Rev(-int - 2) | |
145 | } |
|
145 | } | |
146 | } |
|
146 | } | |
147 | } |
|
147 | } | |
148 |
|
148 | |||
149 | impl From<Element> for RawElement { |
|
149 | impl From<Element> for RawElement { | |
150 | fn from(element: Element) -> RawElement { |
|
150 | fn from(element: Element) -> RawElement { | |
151 | RawElement::from(match element { |
|
151 | RawElement::from(match element { | |
152 | Element::None => 0, |
|
152 | Element::None => 0, | |
153 | Element::Block(i) => i as i32, |
|
153 | Element::Block(i) => i as i32, | |
154 | Element::Rev(rev) => -rev - 2, |
|
154 | Element::Rev(rev) => -rev - 2, | |
155 | }) |
|
155 | }) | |
156 | } |
|
156 | } | |
157 | } |
|
157 | } | |
158 |
|
158 | |||
159 | /// A logical block of the `NodeTree`, packed with a fixed size. |
|
159 | /// A logical block of the `NodeTree`, packed with a fixed size. | |
160 | /// |
|
160 | /// | |
161 | /// These are always used in container types implementing `Index<Block>`, |
|
161 | /// These are always used in container types implementing `Index<Block>`, | |
162 | /// such as `&Block` |
|
162 | /// such as `&Block` | |
163 | /// |
|
163 | /// | |
164 | /// As an array of integers, its ith element encodes that the |
|
164 | /// As an array of integers, its ith element encodes that the | |
165 | /// ith potential edge from the block, representing the ith hexadecimal digit |
|
165 | /// ith potential edge from the block, representing the ith hexadecimal digit | |
166 | /// (nybble) `i` is either: |
|
166 | /// (nybble) `i` is either: | |
167 | /// |
|
167 | /// | |
168 | /// - absent (value -1) |
|
168 | /// - absent (value -1) | |
169 | /// - another `Block` in the same indexable container (value β₯ 0) |
|
169 | /// - another `Block` in the same indexable container (value β₯ 0) | |
170 | /// - a `Revision` leaf (value β€ -2) |
|
170 | /// - a `Revision` leaf (value β€ -2) | |
171 | /// |
|
171 | /// | |
172 | /// Endianness has to be fixed for consistency on shared storage across |
|
172 | /// Endianness has to be fixed for consistency on shared storage across | |
173 | /// different architectures. |
|
173 | /// different architectures. | |
174 | /// |
|
174 | /// | |
175 | /// A key difference with the C `nodetree` is that we need to be |
|
175 | /// A key difference with the C `nodetree` is that we need to be | |
176 | /// able to represent the [`Block`] at index 0, hence -1 is the empty marker |
|
176 | /// able to represent the [`Block`] at index 0, hence -1 is the empty marker | |
177 | /// rather than 0 and the `Revision` range upper limit of -2 instead of -1. |
|
177 | /// rather than 0 and the `Revision` range upper limit of -2 instead of -1. | |
178 | /// |
|
178 | /// | |
179 | /// Another related difference is that `NULL_REVISION` (-1) is not |
|
179 | /// Another related difference is that `NULL_REVISION` (-1) is not | |
180 | /// represented at all, because we want an immutable empty nodetree |
|
180 | /// represented at all, because we want an immutable empty nodetree | |
181 | /// to be valid. |
|
181 | /// to be valid. | |
182 |
|
182 | |||
183 | const ELEMENTS_PER_BLOCK: usize = 16; // number of different values in a nybble |
|
183 | const ELEMENTS_PER_BLOCK: usize = 16; // number of different values in a nybble | |
184 |
|
184 | |||
185 | #[derive(Copy, Clone, BytesCast, PartialEq)] |
|
185 | #[derive(Copy, Clone, BytesCast, PartialEq)] | |
186 | #[repr(transparent)] |
|
186 | #[repr(transparent)] | |
187 | pub struct Block([RawElement; ELEMENTS_PER_BLOCK]); |
|
187 | pub struct Block([RawElement; ELEMENTS_PER_BLOCK]); | |
188 |
|
188 | |||
189 | impl Block { |
|
189 | impl Block { | |
190 | fn new() -> Self { |
|
190 | fn new() -> Self { | |
191 | let absent_node = RawElement::from(-1); |
|
191 | let absent_node = RawElement::from(-1); | |
192 | Block([absent_node; ELEMENTS_PER_BLOCK]) |
|
192 | Block([absent_node; ELEMENTS_PER_BLOCK]) | |
193 | } |
|
193 | } | |
194 |
|
194 | |||
195 | fn get(&self, nybble: u8) -> Element { |
|
195 | fn get(&self, nybble: u8) -> Element { | |
196 | self.0[nybble as usize].into() |
|
196 | self.0[nybble as usize].into() | |
197 | } |
|
197 | } | |
198 |
|
198 | |||
199 | fn set(&mut self, nybble: u8, element: Element) { |
|
199 | fn set(&mut self, nybble: u8, element: Element) { | |
200 | self.0[nybble as usize] = element.into() |
|
200 | self.0[nybble as usize] = element.into() | |
201 | } |
|
201 | } | |
202 | } |
|
202 | } | |
203 |
|
203 | |||
204 | impl fmt::Debug for Block { |
|
204 | impl fmt::Debug for Block { | |
205 | /// sparse representation for testing and debugging purposes |
|
205 | /// sparse representation for testing and debugging purposes | |
206 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { |
|
206 | fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result { | |
207 | f.debug_map() |
|
207 | f.debug_map() | |
208 | .entries((0..16).filter_map(|i| match self.get(i) { |
|
208 | .entries((0..16).filter_map(|i| match self.get(i) { | |
209 | Element::None => None, |
|
209 | Element::None => None, | |
210 | element => Some((i, element)), |
|
210 | element => Some((i, element)), | |
211 | })) |
|
211 | })) | |
212 | .finish() |
|
212 | .finish() | |
213 | } |
|
213 | } | |
214 | } |
|
214 | } | |
215 |
|
215 | |||
216 | /// A mutable 16-radix tree with the root block logically at the end |
|
216 | /// A mutable 16-radix tree with the root block logically at the end | |
217 | /// |
|
217 | /// | |
218 | /// Because of the append only nature of our node trees, we need to |
|
218 | /// Because of the append only nature of our node trees, we need to | |
219 | /// keep the original untouched and store new blocks separately. |
|
219 | /// keep the original untouched and store new blocks separately. | |
220 | /// |
|
220 | /// | |
221 | /// The mutable root `Block` is kept apart so that we don't have to rebump |
|
221 | /// The mutable root `Block` is kept apart so that we don't have to rebump | |
222 | /// it on each insertion. |
|
222 | /// it on each insertion. | |
223 | pub struct NodeTree { |
|
223 | pub struct NodeTree { | |
224 | readonly: Box<dyn Deref<Target = [Block]> + Send>, |
|
224 | readonly: Box<dyn Deref<Target = [Block]> + Send>, | |
225 | growable: Vec<Block>, |
|
225 | growable: Vec<Block>, | |
226 | root: Block, |
|
226 | root: Block, | |
227 | masked_inner_blocks: usize, |
|
227 | masked_inner_blocks: usize, | |
228 | } |
|
228 | } | |
229 |
|
229 | |||
230 | impl Index<usize> for NodeTree { |
|
230 | impl Index<usize> for NodeTree { | |
231 | type Output = Block; |
|
231 | type Output = Block; | |
232 |
|
232 | |||
233 | fn index(&self, i: usize) -> &Block { |
|
233 | fn index(&self, i: usize) -> &Block { | |
234 | let ro_len = self.readonly.len(); |
|
234 | let ro_len = self.readonly.len(); | |
235 | if i < ro_len { |
|
235 | if i < ro_len { | |
236 | &self.readonly[i] |
|
236 | &self.readonly[i] | |
237 | } else if i == ro_len + self.growable.len() { |
|
237 | } else if i == ro_len + self.growable.len() { | |
238 | &self.root |
|
238 | &self.root | |
239 | } else { |
|
239 | } else { | |
240 | &self.growable[i - ro_len] |
|
240 | &self.growable[i - ro_len] | |
241 | } |
|
241 | } | |
242 | } |
|
242 | } | |
243 | } |
|
243 | } | |
244 |
|
244 | |||
245 | /// Return `None` unless the `Node` for `rev` has given prefix in `index`. |
|
245 | /// Return `None` unless the `Node` for `rev` has given prefix in `index`. | |
246 | fn has_prefix_or_none( |
|
246 | fn has_prefix_or_none( | |
247 | idx: &impl RevlogIndex, |
|
247 | idx: &impl RevlogIndex, | |
248 | prefix: NodePrefix, |
|
248 | prefix: NodePrefix, | |
249 | rev: Revision, |
|
249 | rev: Revision, | |
250 | ) -> Result<Option<Revision>, NodeMapError> { |
|
250 | ) -> Result<Option<Revision>, NodeMapError> { | |
251 | idx.node(rev) |
|
251 | idx.node(rev) | |
252 | .ok_or_else(|| NodeMapError::RevisionNotInIndex(rev)) |
|
252 | .ok_or_else(|| NodeMapError::RevisionNotInIndex(rev)) | |
253 | .map(|node| { |
|
253 | .map(|node| { | |
254 | if prefix.is_prefix_of(node) { |
|
254 | if prefix.is_prefix_of(node) { | |
255 | Some(rev) |
|
255 | Some(rev) | |
256 | } else { |
|
256 | } else { | |
257 | None |
|
257 | None | |
258 | } |
|
258 | } | |
259 | }) |
|
259 | }) | |
260 | } |
|
260 | } | |
261 |
|
261 | |||
262 | /// validate that the candidate's node starts indeed with given prefix, |
|
262 | /// validate that the candidate's node starts indeed with given prefix, | |
263 | /// and treat ambiguities related to `NULL_REVISION`. |
|
263 | /// and treat ambiguities related to `NULL_REVISION`. | |
264 | /// |
|
264 | /// | |
265 | /// From the data in the NodeTree, one can only conclude that some |
|
265 | /// From the data in the NodeTree, one can only conclude that some | |
266 | /// revision is the only one for a *subprefix* of the one being looked up. |
|
266 | /// revision is the only one for a *subprefix* of the one being looked up. | |
267 | fn validate_candidate( |
|
267 | fn validate_candidate( | |
268 | idx: &impl RevlogIndex, |
|
268 | idx: &impl RevlogIndex, | |
269 | prefix: NodePrefix, |
|
269 | prefix: NodePrefix, | |
270 | candidate: (Option<Revision>, usize), |
|
270 | candidate: (Option<Revision>, usize), | |
271 | ) -> Result<(Option<Revision>, usize), NodeMapError> { |
|
271 | ) -> Result<(Option<Revision>, usize), NodeMapError> { | |
272 | let (rev, steps) = candidate; |
|
272 | let (rev, steps) = candidate; | |
273 | if let Some(nz_nybble) = prefix.first_different_nybble(&NULL_NODE) { |
|
273 | if let Some(nz_nybble) = prefix.first_different_nybble(&NULL_NODE) { | |
274 | rev.map_or(Ok((None, steps)), |r| { |
|
274 | rev.map_or(Ok((None, steps)), |r| { | |
275 | has_prefix_or_none(idx, prefix, r) |
|
275 | has_prefix_or_none(idx, prefix, r) | |
276 | .map(|opt| (opt, max(steps, nz_nybble + 1))) |
|
276 | .map(|opt| (opt, max(steps, nz_nybble + 1))) | |
277 | }) |
|
277 | }) | |
278 | } else { |
|
278 | } else { | |
279 | // the prefix is only made of zeros; NULL_REVISION always matches it |
|
279 | // the prefix is only made of zeros; NULL_REVISION always matches it | |
280 | // and any other *valid* result is an ambiguity |
|
280 | // and any other *valid* result is an ambiguity | |
281 | match rev { |
|
281 | match rev { | |
282 | None => Ok((Some(NULL_REVISION), steps + 1)), |
|
282 | None => Ok((Some(NULL_REVISION), steps + 1)), | |
283 | Some(r) => match has_prefix_or_none(idx, prefix, r)? { |
|
283 | Some(r) => match has_prefix_or_none(idx, prefix, r)? { | |
284 | None => Ok((Some(NULL_REVISION), steps + 1)), |
|
284 | None => Ok((Some(NULL_REVISION), steps + 1)), | |
285 | _ => Err(NodeMapError::MultipleResults), |
|
285 | _ => Err(NodeMapError::MultipleResults), | |
286 | }, |
|
286 | }, | |
287 | } |
|
287 | } | |
288 | } |
|
288 | } | |
289 | } |
|
289 | } | |
290 |
|
290 | |||
291 | impl NodeTree { |
|
291 | impl NodeTree { | |
292 | /// Initiate a NodeTree from an immutable slice-like of `Block` |
|
292 | /// Initiate a NodeTree from an immutable slice-like of `Block` | |
293 | /// |
|
293 | /// | |
294 | /// We keep `readonly` and clone its root block if it isn't empty. |
|
294 | /// We keep `readonly` and clone its root block if it isn't empty. | |
295 | fn new(readonly: Box<dyn Deref<Target = [Block]> + Send>) -> Self { |
|
295 | fn new(readonly: Box<dyn Deref<Target = [Block]> + Send>) -> Self { | |
296 | let root = readonly.last().cloned().unwrap_or_else(Block::new); |
|
296 | let root = readonly.last().cloned().unwrap_or_else(Block::new); | |
297 | NodeTree { |
|
297 | NodeTree { | |
298 | readonly, |
|
298 | readonly, | |
299 | growable: Vec::new(), |
|
299 | growable: Vec::new(), | |
300 | root, |
|
300 | root, | |
301 | masked_inner_blocks: 0, |
|
301 | masked_inner_blocks: 0, | |
302 | } |
|
302 | } | |
303 | } |
|
303 | } | |
304 |
|
304 | |||
305 | /// Create from an opaque bunch of bytes |
|
305 | /// Create from an opaque bunch of bytes | |
306 | /// |
|
306 | /// | |
307 | /// The created `NodeTreeBytes` from `buffer`, |
|
307 | /// The created `NodeTreeBytes` from `buffer`, | |
308 | /// of which exactly `amount` bytes are used. |
|
308 | /// of which exactly `amount` bytes are used. | |
309 | /// |
|
309 | /// | |
310 | /// - `buffer` could be derived from `PyBuffer` and `Mmap` objects. |
|
310 | /// - `buffer` could be derived from `PyBuffer` and `Mmap` objects. | |
311 | /// - `offset` allows for the final file format to include fixed data |
|
311 | /// - `offset` allows for the final file format to include fixed data | |
312 | /// (generation number, behavioural flags) |
|
312 | /// (generation number, behavioural flags) | |
313 | /// - `amount` is expressed in bytes, and is not automatically derived from |
|
313 | /// - `amount` is expressed in bytes, and is not automatically derived from | |
314 | /// `bytes`, so that a caller that manages them atomically can perform |
|
314 | /// `bytes`, so that a caller that manages them atomically can perform | |
315 | /// temporary disk serializations and still rollback easily if needed. |
|
315 | /// temporary disk serializations and still rollback easily if needed. | |
316 | /// First use-case for this would be to support Mercurial shell hooks. |
|
316 | /// First use-case for this would be to support Mercurial shell hooks. | |
317 | /// |
|
317 | /// | |
318 | /// panics if `buffer` is smaller than `amount` |
|
318 | /// panics if `buffer` is smaller than `amount` | |
319 | pub fn load_bytes( |
|
319 | pub fn load_bytes( | |
320 | bytes: Box<dyn Deref<Target = [u8]> + Send>, |
|
320 | bytes: Box<dyn Deref<Target = [u8]> + Send>, | |
321 | amount: usize, |
|
321 | amount: usize, | |
322 | ) -> Self { |
|
322 | ) -> Self { | |
323 | NodeTree::new(Box::new(NodeTreeBytes::new(bytes, amount))) |
|
323 | NodeTree::new(Box::new(NodeTreeBytes::new(bytes, amount))) | |
324 | } |
|
324 | } | |
325 |
|
325 | |||
326 | /// Retrieve added `Block` and the original immutable data |
|
326 | /// Retrieve added `Block` and the original immutable data | |
327 | pub fn into_readonly_and_added( |
|
327 | pub fn into_readonly_and_added( | |
328 | self, |
|
328 | self, | |
329 | ) -> (Box<dyn Deref<Target = [Block]> + Send>, Vec<Block>) { |
|
329 | ) -> (Box<dyn Deref<Target = [Block]> + Send>, Vec<Block>) { | |
330 | let mut vec = self.growable; |
|
330 | let mut vec = self.growable; | |
331 | let readonly = self.readonly; |
|
331 | let readonly = self.readonly; | |
332 | if readonly.last() != Some(&self.root) { |
|
332 | if readonly.last() != Some(&self.root) { | |
333 | vec.push(self.root); |
|
333 | vec.push(self.root); | |
334 | } |
|
334 | } | |
335 | (readonly, vec) |
|
335 | (readonly, vec) | |
336 | } |
|
336 | } | |
337 |
|
337 | |||
338 | /// Retrieve added `Blocks` as bytes, ready to be written to persistent |
|
338 | /// Retrieve added `Blocks` as bytes, ready to be written to persistent | |
339 | /// storage |
|
339 | /// storage | |
340 | pub fn into_readonly_and_added_bytes( |
|
340 | pub fn into_readonly_and_added_bytes( | |
341 | self, |
|
341 | self, | |
342 | ) -> (Box<dyn Deref<Target = [Block]> + Send>, Vec<u8>) { |
|
342 | ) -> (Box<dyn Deref<Target = [Block]> + Send>, Vec<u8>) { | |
343 | let (readonly, vec) = self.into_readonly_and_added(); |
|
343 | let (readonly, vec) = self.into_readonly_and_added(); | |
344 | // Prevent running `v`'s destructor so we are in complete control |
|
344 | // Prevent running `v`'s destructor so we are in complete control | |
345 | // of the allocation. |
|
345 | // of the allocation. | |
346 | let vec = mem::ManuallyDrop::new(vec); |
|
346 | let vec = mem::ManuallyDrop::new(vec); | |
347 |
|
347 | |||
348 | // Transmute the `Vec<Block>` to a `Vec<u8>`. Blocks are contiguous |
|
348 | // Transmute the `Vec<Block>` to a `Vec<u8>`. Blocks are contiguous | |
349 | // bytes, so this is perfectly safe. |
|
349 | // bytes, so this is perfectly safe. | |
350 | let bytes = unsafe { |
|
350 | let bytes = unsafe { | |
351 | // Check for compatible allocation layout. |
|
351 | // Check for compatible allocation layout. | |
352 | // (Optimized away by constant-folding + dead code elimination.) |
|
352 | // (Optimized away by constant-folding + dead code elimination.) | |
353 | assert_eq!(size_of::<Block>(), 64); |
|
353 | assert_eq!(size_of::<Block>(), 64); | |
354 | assert_eq!(align_of::<Block>(), 1); |
|
354 | assert_eq!(align_of::<Block>(), 1); | |
355 |
|
355 | |||
356 | // /!\ Any use of `vec` after this is use-after-free. |
|
356 | // /!\ Any use of `vec` after this is use-after-free. | |
357 | // TODO: use `into_raw_parts` once stabilized |
|
357 | // TODO: use `into_raw_parts` once stabilized | |
358 | Vec::from_raw_parts( |
|
358 | Vec::from_raw_parts( | |
359 | vec.as_ptr() as *mut u8, |
|
359 | vec.as_ptr() as *mut u8, | |
360 | vec.len() * size_of::<Block>(), |
|
360 | vec.len() * size_of::<Block>(), | |
361 | vec.capacity() * size_of::<Block>(), |
|
361 | vec.capacity() * size_of::<Block>(), | |
362 | ) |
|
362 | ) | |
363 | }; |
|
363 | }; | |
364 | (readonly, bytes) |
|
364 | (readonly, bytes) | |
365 | } |
|
365 | } | |
366 |
|
366 | |||
367 | /// Total number of blocks |
|
367 | /// Total number of blocks | |
368 | fn len(&self) -> usize { |
|
368 | fn len(&self) -> usize { | |
369 | self.readonly.len() + self.growable.len() + 1 |
|
369 | self.readonly.len() + self.growable.len() + 1 | |
370 | } |
|
370 | } | |
371 |
|
371 | |||
372 | /// Implemented for completeness |
|
372 | /// Implemented for completeness | |
373 | /// |
|
373 | /// | |
374 | /// A `NodeTree` always has at least the mutable root block. |
|
374 | /// A `NodeTree` always has at least the mutable root block. | |
375 | #[allow(dead_code)] |
|
375 | #[allow(dead_code)] | |
376 | fn is_empty(&self) -> bool { |
|
376 | fn is_empty(&self) -> bool { | |
377 | false |
|
377 | false | |
378 | } |
|
378 | } | |
379 |
|
379 | |||
380 | /// Main working method for `NodeTree` searches |
|
380 | /// Main working method for `NodeTree` searches | |
381 | /// |
|
381 | /// | |
382 | /// The first returned value is the result of analysing `NodeTree` data |
|
382 | /// The first returned value is the result of analysing `NodeTree` data | |
383 | /// *alone*: whereas `None` guarantees that the given prefix is absent |
|
383 | /// *alone*: whereas `None` guarantees that the given prefix is absent | |
384 | /// from the `NodeTree` data (but still could match `NULL_NODE`), with |
|
384 | /// from the `NodeTree` data (but still could match `NULL_NODE`), with | |
385 | /// `Some(rev)`, it is to be understood that `rev` is the unique `Revision` |
|
385 | /// `Some(rev)`, it is to be understood that `rev` is the unique `Revision` | |
386 | /// that could match the prefix. Actually, all that can be inferred from |
|
386 | /// that could match the prefix. Actually, all that can be inferred from | |
387 | /// the `NodeTree` data is that `rev` is the revision with the longest |
|
387 | /// the `NodeTree` data is that `rev` is the revision with the longest | |
388 | /// common node prefix with the given prefix. |
|
388 | /// common node prefix with the given prefix. | |
389 | /// |
|
389 | /// | |
390 | /// The second returned value is the size of the smallest subprefix |
|
390 | /// The second returned value is the size of the smallest subprefix | |
391 | /// of `prefix` that would give the same result, i.e. not the |
|
391 | /// of `prefix` that would give the same result, i.e. not the | |
392 | /// `MultipleResults` error variant (again, using only the data of the |
|
392 | /// `MultipleResults` error variant (again, using only the data of the | |
393 | /// `NodeTree`). |
|
393 | /// `NodeTree`). | |
394 | fn lookup( |
|
394 | fn lookup( | |
395 | &self, |
|
395 | &self, | |
396 | prefix: NodePrefix, |
|
396 | prefix: NodePrefix, | |
397 | ) -> Result<(Option<Revision>, usize), NodeMapError> { |
|
397 | ) -> Result<(Option<Revision>, usize), NodeMapError> { | |
398 | for (i, visit_item) in self.visit(prefix).enumerate() { |
|
398 | for (i, visit_item) in self.visit(prefix).enumerate() { | |
399 | if let Some(opt) = visit_item.final_revision() { |
|
399 | if let Some(opt) = visit_item.final_revision() { | |
400 | return Ok((opt, i + 1)); |
|
400 | return Ok((opt, i + 1)); | |
401 | } |
|
401 | } | |
402 | } |
|
402 | } | |
403 | Err(NodeMapError::MultipleResults) |
|
403 | Err(NodeMapError::MultipleResults) | |
404 | } |
|
404 | } | |
405 |
|
405 | |||
406 |
fn visit |
|
406 | fn visit(&self, prefix: NodePrefix) -> NodeTreeVisitor { | |
407 | NodeTreeVisitor { |
|
407 | NodeTreeVisitor { | |
408 | nt: self, |
|
408 | nt: self, | |
409 | prefix, |
|
409 | prefix, | |
410 | visit: self.len() - 1, |
|
410 | visit: self.len() - 1, | |
411 | nybble_idx: 0, |
|
411 | nybble_idx: 0, | |
412 | done: false, |
|
412 | done: false, | |
413 | } |
|
413 | } | |
414 | } |
|
414 | } | |
415 | /// Return a mutable reference for `Block` at index `idx`. |
|
415 | /// Return a mutable reference for `Block` at index `idx`. | |
416 | /// |
|
416 | /// | |
417 | /// If `idx` lies in the immutable area, then the reference is to |
|
417 | /// If `idx` lies in the immutable area, then the reference is to | |
418 | /// a newly appended copy. |
|
418 | /// a newly appended copy. | |
419 | /// |
|
419 | /// | |
420 | /// Returns (new_idx, glen, mut_ref) where |
|
420 | /// Returns (new_idx, glen, mut_ref) where | |
421 | /// |
|
421 | /// | |
422 | /// - `new_idx` is the index of the mutable `Block` |
|
422 | /// - `new_idx` is the index of the mutable `Block` | |
423 | /// - `mut_ref` is a mutable reference to the mutable Block. |
|
423 | /// - `mut_ref` is a mutable reference to the mutable Block. | |
424 | /// - `glen` is the new length of `self.growable` |
|
424 | /// - `glen` is the new length of `self.growable` | |
425 | /// |
|
425 | /// | |
426 | /// Note: the caller wouldn't be allowed to query `self.growable.len()` |
|
426 | /// Note: the caller wouldn't be allowed to query `self.growable.len()` | |
427 | /// itself because of the mutable borrow taken with the returned `Block` |
|
427 | /// itself because of the mutable borrow taken with the returned `Block` | |
428 | fn mutable_block(&mut self, idx: usize) -> (usize, &mut Block, usize) { |
|
428 | fn mutable_block(&mut self, idx: usize) -> (usize, &mut Block, usize) { | |
429 | let ro_blocks = &self.readonly; |
|
429 | let ro_blocks = &self.readonly; | |
430 | let ro_len = ro_blocks.len(); |
|
430 | let ro_len = ro_blocks.len(); | |
431 | let glen = self.growable.len(); |
|
431 | let glen = self.growable.len(); | |
432 | if idx < ro_len { |
|
432 | if idx < ro_len { | |
433 | self.masked_inner_blocks += 1; |
|
433 | self.masked_inner_blocks += 1; | |
434 | self.growable.push(ro_blocks[idx]); |
|
434 | self.growable.push(ro_blocks[idx]); | |
435 | (glen + ro_len, &mut self.growable[glen], glen + 1) |
|
435 | (glen + ro_len, &mut self.growable[glen], glen + 1) | |
436 | } else if glen + ro_len == idx { |
|
436 | } else if glen + ro_len == idx { | |
437 | (idx, &mut self.root, glen) |
|
437 | (idx, &mut self.root, glen) | |
438 | } else { |
|
438 | } else { | |
439 | (idx, &mut self.growable[idx - ro_len], glen) |
|
439 | (idx, &mut self.growable[idx - ro_len], glen) | |
440 | } |
|
440 | } | |
441 | } |
|
441 | } | |
442 |
|
442 | |||
443 | /// Main insertion method |
|
443 | /// Main insertion method | |
444 | /// |
|
444 | /// | |
445 | /// This will dive in the node tree to find the deepest `Block` for |
|
445 | /// This will dive in the node tree to find the deepest `Block` for | |
446 | /// `node`, split it as much as needed and record `node` in there. |
|
446 | /// `node`, split it as much as needed and record `node` in there. | |
447 | /// The method then backtracks, updating references in all the visited |
|
447 | /// The method then backtracks, updating references in all the visited | |
448 | /// blocks from the root. |
|
448 | /// blocks from the root. | |
449 | /// |
|
449 | /// | |
450 | /// All the mutated `Block` are copied first to the growable part if |
|
450 | /// All the mutated `Block` are copied first to the growable part if | |
451 | /// needed. That happens for those in the immutable part except the root. |
|
451 | /// needed. That happens for those in the immutable part except the root. | |
452 | pub fn insert<I: RevlogIndex>( |
|
452 | pub fn insert<I: RevlogIndex>( | |
453 | &mut self, |
|
453 | &mut self, | |
454 | index: &I, |
|
454 | index: &I, | |
455 | node: &Node, |
|
455 | node: &Node, | |
456 | rev: Revision, |
|
456 | rev: Revision, | |
457 | ) -> Result<(), NodeMapError> { |
|
457 | ) -> Result<(), NodeMapError> { | |
458 | let ro_len = &self.readonly.len(); |
|
458 | let ro_len = &self.readonly.len(); | |
459 |
|
459 | |||
460 | let mut visit_steps: Vec<_> = self.visit(node.into()).collect(); |
|
460 | let mut visit_steps: Vec<_> = self.visit(node.into()).collect(); | |
461 | let read_nybbles = visit_steps.len(); |
|
461 | let read_nybbles = visit_steps.len(); | |
462 | // visit_steps cannot be empty, since we always visit the root block |
|
462 | // visit_steps cannot be empty, since we always visit the root block | |
463 | let deepest = visit_steps.pop().unwrap(); |
|
463 | let deepest = visit_steps.pop().unwrap(); | |
464 |
|
464 | |||
465 | let (mut block_idx, mut block, mut glen) = |
|
465 | let (mut block_idx, mut block, mut glen) = | |
466 | self.mutable_block(deepest.block_idx); |
|
466 | self.mutable_block(deepest.block_idx); | |
467 |
|
467 | |||
468 | if let Element::Rev(old_rev) = deepest.element { |
|
468 | if let Element::Rev(old_rev) = deepest.element { | |
469 | let old_node = index |
|
469 | let old_node = index | |
470 | .node(old_rev) |
|
470 | .node(old_rev) | |
471 | .ok_or_else(|| NodeMapError::RevisionNotInIndex(old_rev))?; |
|
471 | .ok_or_else(|| NodeMapError::RevisionNotInIndex(old_rev))?; | |
472 | if old_node == node { |
|
472 | if old_node == node { | |
473 | return Ok(()); // avoid creating lots of useless blocks |
|
473 | return Ok(()); // avoid creating lots of useless blocks | |
474 | } |
|
474 | } | |
475 |
|
475 | |||
476 | // Looping over the tail of nybbles in both nodes, creating |
|
476 | // Looping over the tail of nybbles in both nodes, creating | |
477 | // new blocks until we find the difference |
|
477 | // new blocks until we find the difference | |
478 | let mut new_block_idx = ro_len + glen; |
|
478 | let mut new_block_idx = ro_len + glen; | |
479 | let mut nybble = deepest.nybble; |
|
479 | let mut nybble = deepest.nybble; | |
480 | for nybble_pos in read_nybbles..node.nybbles_len() { |
|
480 | for nybble_pos in read_nybbles..node.nybbles_len() { | |
481 | block.set(nybble, Element::Block(new_block_idx)); |
|
481 | block.set(nybble, Element::Block(new_block_idx)); | |
482 |
|
482 | |||
483 | let new_nybble = node.get_nybble(nybble_pos); |
|
483 | let new_nybble = node.get_nybble(nybble_pos); | |
484 | let old_nybble = old_node.get_nybble(nybble_pos); |
|
484 | let old_nybble = old_node.get_nybble(nybble_pos); | |
485 |
|
485 | |||
486 | if old_nybble == new_nybble { |
|
486 | if old_nybble == new_nybble { | |
487 | self.growable.push(Block::new()); |
|
487 | self.growable.push(Block::new()); | |
488 | block = &mut self.growable[glen]; |
|
488 | block = &mut self.growable[glen]; | |
489 | glen += 1; |
|
489 | glen += 1; | |
490 | new_block_idx += 1; |
|
490 | new_block_idx += 1; | |
491 | nybble = new_nybble; |
|
491 | nybble = new_nybble; | |
492 | } else { |
|
492 | } else { | |
493 | let mut new_block = Block::new(); |
|
493 | let mut new_block = Block::new(); | |
494 | new_block.set(old_nybble, Element::Rev(old_rev)); |
|
494 | new_block.set(old_nybble, Element::Rev(old_rev)); | |
495 | new_block.set(new_nybble, Element::Rev(rev)); |
|
495 | new_block.set(new_nybble, Element::Rev(rev)); | |
496 | self.growable.push(new_block); |
|
496 | self.growable.push(new_block); | |
497 | break; |
|
497 | break; | |
498 | } |
|
498 | } | |
499 | } |
|
499 | } | |
500 | } else { |
|
500 | } else { | |
501 | // Free slot in the deepest block: no splitting has to be done |
|
501 | // Free slot in the deepest block: no splitting has to be done | |
502 | block.set(deepest.nybble, Element::Rev(rev)); |
|
502 | block.set(deepest.nybble, Element::Rev(rev)); | |
503 | } |
|
503 | } | |
504 |
|
504 | |||
505 | // Backtrack over visit steps to update references |
|
505 | // Backtrack over visit steps to update references | |
506 | while let Some(visited) = visit_steps.pop() { |
|
506 | while let Some(visited) = visit_steps.pop() { | |
507 | let to_write = Element::Block(block_idx); |
|
507 | let to_write = Element::Block(block_idx); | |
508 | if visit_steps.is_empty() { |
|
508 | if visit_steps.is_empty() { | |
509 | self.root.set(visited.nybble, to_write); |
|
509 | self.root.set(visited.nybble, to_write); | |
510 | break; |
|
510 | break; | |
511 | } |
|
511 | } | |
512 | let (new_idx, block, _) = self.mutable_block(visited.block_idx); |
|
512 | let (new_idx, block, _) = self.mutable_block(visited.block_idx); | |
513 | if block.get(visited.nybble) == to_write { |
|
513 | if block.get(visited.nybble) == to_write { | |
514 | break; |
|
514 | break; | |
515 | } |
|
515 | } | |
516 | block.set(visited.nybble, to_write); |
|
516 | block.set(visited.nybble, to_write); | |
517 | block_idx = new_idx; |
|
517 | block_idx = new_idx; | |
518 | } |
|
518 | } | |
519 | Ok(()) |
|
519 | Ok(()) | |
520 | } |
|
520 | } | |
521 |
|
521 | |||
522 | /// Make the whole `NodeTree` logically empty, without touching the |
|
522 | /// Make the whole `NodeTree` logically empty, without touching the | |
523 | /// immutable part. |
|
523 | /// immutable part. | |
524 | pub fn invalidate_all(&mut self) { |
|
524 | pub fn invalidate_all(&mut self) { | |
525 | self.root = Block::new(); |
|
525 | self.root = Block::new(); | |
526 | self.growable = Vec::new(); |
|
526 | self.growable = Vec::new(); | |
527 | self.masked_inner_blocks = self.readonly.len(); |
|
527 | self.masked_inner_blocks = self.readonly.len(); | |
528 | } |
|
528 | } | |
529 |
|
529 | |||
530 | /// Return the number of blocks in the readonly part that are currently |
|
530 | /// Return the number of blocks in the readonly part that are currently | |
531 | /// masked in the mutable part. |
|
531 | /// masked in the mutable part. | |
532 | /// |
|
532 | /// | |
533 | /// The `NodeTree` structure has no efficient way to know how many blocks |
|
533 | /// The `NodeTree` structure has no efficient way to know how many blocks | |
534 | /// are already unreachable in the readonly part. |
|
534 | /// are already unreachable in the readonly part. | |
535 | /// |
|
535 | /// | |
536 | /// After a call to `invalidate_all()`, the returned number can be actually |
|
536 | /// After a call to `invalidate_all()`, the returned number can be actually | |
537 | /// bigger than the whole readonly part, a conventional way to mean that |
|
537 | /// bigger than the whole readonly part, a conventional way to mean that | |
538 | /// all the readonly blocks have been masked. This is what is really |
|
538 | /// all the readonly blocks have been masked. This is what is really | |
539 | /// useful to the caller and does not require to know how many were |
|
539 | /// useful to the caller and does not require to know how many were | |
540 | /// actually unreachable to begin with. |
|
540 | /// actually unreachable to begin with. | |
541 | pub fn masked_readonly_blocks(&self) -> usize { |
|
541 | pub fn masked_readonly_blocks(&self) -> usize { | |
542 | if let Some(readonly_root) = self.readonly.last() { |
|
542 | if let Some(readonly_root) = self.readonly.last() { | |
543 | if readonly_root == &self.root { |
|
543 | if readonly_root == &self.root { | |
544 | return 0; |
|
544 | return 0; | |
545 | } |
|
545 | } | |
546 | } else { |
|
546 | } else { | |
547 | return 0; |
|
547 | return 0; | |
548 | } |
|
548 | } | |
549 | self.masked_inner_blocks + 1 |
|
549 | self.masked_inner_blocks + 1 | |
550 | } |
|
550 | } | |
551 | } |
|
551 | } | |
552 |
|
552 | |||
553 | pub struct NodeTreeBytes { |
|
553 | pub struct NodeTreeBytes { | |
554 | buffer: Box<dyn Deref<Target = [u8]> + Send>, |
|
554 | buffer: Box<dyn Deref<Target = [u8]> + Send>, | |
555 | len_in_blocks: usize, |
|
555 | len_in_blocks: usize, | |
556 | } |
|
556 | } | |
557 |
|
557 | |||
558 | impl NodeTreeBytes { |
|
558 | impl NodeTreeBytes { | |
559 | fn new( |
|
559 | fn new( | |
560 | buffer: Box<dyn Deref<Target = [u8]> + Send>, |
|
560 | buffer: Box<dyn Deref<Target = [u8]> + Send>, | |
561 | amount: usize, |
|
561 | amount: usize, | |
562 | ) -> Self { |
|
562 | ) -> Self { | |
563 | assert!(buffer.len() >= amount); |
|
563 | assert!(buffer.len() >= amount); | |
564 | let len_in_blocks = amount / size_of::<Block>(); |
|
564 | let len_in_blocks = amount / size_of::<Block>(); | |
565 | NodeTreeBytes { |
|
565 | NodeTreeBytes { | |
566 | buffer, |
|
566 | buffer, | |
567 | len_in_blocks, |
|
567 | len_in_blocks, | |
568 | } |
|
568 | } | |
569 | } |
|
569 | } | |
570 | } |
|
570 | } | |
571 |
|
571 | |||
572 | impl Deref for NodeTreeBytes { |
|
572 | impl Deref for NodeTreeBytes { | |
573 | type Target = [Block]; |
|
573 | type Target = [Block]; | |
574 |
|
574 | |||
575 | fn deref(&self) -> &[Block] { |
|
575 | fn deref(&self) -> &[Block] { | |
576 | Block::slice_from_bytes(&self.buffer, self.len_in_blocks) |
|
576 | Block::slice_from_bytes(&self.buffer, self.len_in_blocks) | |
577 | // `NodeTreeBytes::new` already asserted that `self.buffer` is |
|
577 | // `NodeTreeBytes::new` already asserted that `self.buffer` is | |
578 | // large enough. |
|
578 | // large enough. | |
579 | .unwrap() |
|
579 | .unwrap() | |
580 | .0 |
|
580 | .0 | |
581 | } |
|
581 | } | |
582 | } |
|
582 | } | |
583 |
|
583 | |||
584 | struct NodeTreeVisitor<'n> { |
|
584 | struct NodeTreeVisitor<'n> { | |
585 | nt: &'n NodeTree, |
|
585 | nt: &'n NodeTree, | |
586 | prefix: NodePrefix, |
|
586 | prefix: NodePrefix, | |
587 | visit: usize, |
|
587 | visit: usize, | |
588 | nybble_idx: usize, |
|
588 | nybble_idx: usize, | |
589 | done: bool, |
|
589 | done: bool, | |
590 | } |
|
590 | } | |
591 |
|
591 | |||
592 | #[derive(Debug, PartialEq, Clone)] |
|
592 | #[derive(Debug, PartialEq, Clone)] | |
593 | struct NodeTreeVisitItem { |
|
593 | struct NodeTreeVisitItem { | |
594 | block_idx: usize, |
|
594 | block_idx: usize, | |
595 | nybble: u8, |
|
595 | nybble: u8, | |
596 | element: Element, |
|
596 | element: Element, | |
597 | } |
|
597 | } | |
598 |
|
598 | |||
599 | impl<'n> Iterator for NodeTreeVisitor<'n> { |
|
599 | impl<'n> Iterator for NodeTreeVisitor<'n> { | |
600 | type Item = NodeTreeVisitItem; |
|
600 | type Item = NodeTreeVisitItem; | |
601 |
|
601 | |||
602 | fn next(&mut self) -> Option<Self::Item> { |
|
602 | fn next(&mut self) -> Option<Self::Item> { | |
603 | if self.done || self.nybble_idx >= self.prefix.nybbles_len() { |
|
603 | if self.done || self.nybble_idx >= self.prefix.nybbles_len() { | |
604 | return None; |
|
604 | return None; | |
605 | } |
|
605 | } | |
606 |
|
606 | |||
607 | let nybble = self.prefix.get_nybble(self.nybble_idx); |
|
607 | let nybble = self.prefix.get_nybble(self.nybble_idx); | |
608 | self.nybble_idx += 1; |
|
608 | self.nybble_idx += 1; | |
609 |
|
609 | |||
610 | let visit = self.visit; |
|
610 | let visit = self.visit; | |
611 | let element = self.nt[visit].get(nybble); |
|
611 | let element = self.nt[visit].get(nybble); | |
612 | if let Element::Block(idx) = element { |
|
612 | if let Element::Block(idx) = element { | |
613 | self.visit = idx; |
|
613 | self.visit = idx; | |
614 | } else { |
|
614 | } else { | |
615 | self.done = true; |
|
615 | self.done = true; | |
616 | } |
|
616 | } | |
617 |
|
617 | |||
618 | Some(NodeTreeVisitItem { |
|
618 | Some(NodeTreeVisitItem { | |
619 | block_idx: visit, |
|
619 | block_idx: visit, | |
620 | nybble, |
|
620 | nybble, | |
621 | element, |
|
621 | element, | |
622 | }) |
|
622 | }) | |
623 | } |
|
623 | } | |
624 | } |
|
624 | } | |
625 |
|
625 | |||
626 | impl NodeTreeVisitItem { |
|
626 | impl NodeTreeVisitItem { | |
627 | // Return `Some(opt)` if this item is final, with `opt` being the |
|
627 | // Return `Some(opt)` if this item is final, with `opt` being the | |
628 | // `Revision` that it may represent. |
|
628 | // `Revision` that it may represent. | |
629 | // |
|
629 | // | |
630 | // If the item is not terminal, return `None` |
|
630 | // If the item is not terminal, return `None` | |
631 | fn final_revision(&self) -> Option<Option<Revision>> { |
|
631 | fn final_revision(&self) -> Option<Option<Revision>> { | |
632 | match self.element { |
|
632 | match self.element { | |
633 | Element::Block(_) => None, |
|
633 | Element::Block(_) => None, | |
634 | Element::Rev(r) => Some(Some(r)), |
|
634 | Element::Rev(r) => Some(Some(r)), | |
635 | Element::None => Some(None), |
|
635 | Element::None => Some(None), | |
636 | } |
|
636 | } | |
637 | } |
|
637 | } | |
638 | } |
|
638 | } | |
639 |
|
639 | |||
640 | impl From<Vec<Block>> for NodeTree { |
|
640 | impl From<Vec<Block>> for NodeTree { | |
641 | fn from(vec: Vec<Block>) -> Self { |
|
641 | fn from(vec: Vec<Block>) -> Self { | |
642 | Self::new(Box::new(vec)) |
|
642 | Self::new(Box::new(vec)) | |
643 | } |
|
643 | } | |
644 | } |
|
644 | } | |
645 |
|
645 | |||
646 | impl fmt::Debug for NodeTree { |
|
646 | impl fmt::Debug for NodeTree { | |
647 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { |
|
647 | fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result { | |
648 | let readonly: &[Block] = &*self.readonly; |
|
648 | let readonly: &[Block] = &*self.readonly; | |
649 | write!( |
|
649 | write!( | |
650 | f, |
|
650 | f, | |
651 | "readonly: {:?}, growable: {:?}, root: {:?}", |
|
651 | "readonly: {:?}, growable: {:?}, root: {:?}", | |
652 | readonly, self.growable, self.root |
|
652 | readonly, self.growable, self.root | |
653 | ) |
|
653 | ) | |
654 | } |
|
654 | } | |
655 | } |
|
655 | } | |
656 |
|
656 | |||
657 | impl Default for NodeTree { |
|
657 | impl Default for NodeTree { | |
658 | /// Create a fully mutable empty NodeTree |
|
658 | /// Create a fully mutable empty NodeTree | |
659 | fn default() -> Self { |
|
659 | fn default() -> Self { | |
660 | NodeTree::new(Box::new(Vec::new())) |
|
660 | NodeTree::new(Box::new(Vec::new())) | |
661 | } |
|
661 | } | |
662 | } |
|
662 | } | |
663 |
|
663 | |||
664 | impl NodeMap for NodeTree { |
|
664 | impl NodeMap for NodeTree { | |
665 | fn find_bin<'a>( |
|
665 | fn find_bin<'a>( | |
666 | &self, |
|
666 | &self, | |
667 | idx: &impl RevlogIndex, |
|
667 | idx: &impl RevlogIndex, | |
668 | prefix: NodePrefix, |
|
668 | prefix: NodePrefix, | |
669 | ) -> Result<Option<Revision>, NodeMapError> { |
|
669 | ) -> Result<Option<Revision>, NodeMapError> { | |
670 | validate_candidate(idx, prefix, self.lookup(prefix)?) |
|
670 | validate_candidate(idx, prefix, self.lookup(prefix)?) | |
671 | .map(|(opt, _shortest)| opt) |
|
671 | .map(|(opt, _shortest)| opt) | |
672 | } |
|
672 | } | |
673 |
|
673 | |||
674 | fn unique_prefix_len_bin<'a>( |
|
674 | fn unique_prefix_len_bin<'a>( | |
675 | &self, |
|
675 | &self, | |
676 | idx: &impl RevlogIndex, |
|
676 | idx: &impl RevlogIndex, | |
677 | prefix: NodePrefix, |
|
677 | prefix: NodePrefix, | |
678 | ) -> Result<Option<usize>, NodeMapError> { |
|
678 | ) -> Result<Option<usize>, NodeMapError> { | |
679 | validate_candidate(idx, prefix, self.lookup(prefix)?) |
|
679 | validate_candidate(idx, prefix, self.lookup(prefix)?) | |
680 | .map(|(opt, shortest)| opt.map(|_rev| shortest)) |
|
680 | .map(|(opt, shortest)| opt.map(|_rev| shortest)) | |
681 | } |
|
681 | } | |
682 | } |
|
682 | } | |
683 |
|
683 | |||
684 | #[cfg(test)] |
|
684 | #[cfg(test)] | |
685 | mod tests { |
|
685 | mod tests { | |
686 | use super::NodeMapError::*; |
|
686 | use super::NodeMapError::*; | |
687 | use super::*; |
|
687 | use super::*; | |
688 | use crate::revlog::node::{hex_pad_right, Node}; |
|
688 | use crate::revlog::node::{hex_pad_right, Node}; | |
689 | use std::collections::HashMap; |
|
689 | use std::collections::HashMap; | |
690 |
|
690 | |||
691 | /// Creates a `Block` using a syntax close to the `Debug` output |
|
691 | /// Creates a `Block` using a syntax close to the `Debug` output | |
692 | macro_rules! block { |
|
692 | macro_rules! block { | |
693 | {$($nybble:tt : $variant:ident($val:tt)),*} => ( |
|
693 | {$($nybble:tt : $variant:ident($val:tt)),*} => ( | |
694 | { |
|
694 | { | |
695 | let mut block = Block::new(); |
|
695 | let mut block = Block::new(); | |
696 | $(block.set($nybble, Element::$variant($val)));*; |
|
696 | $(block.set($nybble, Element::$variant($val)));*; | |
697 | block |
|
697 | block | |
698 | } |
|
698 | } | |
699 | ) |
|
699 | ) | |
700 | } |
|
700 | } | |
701 |
|
701 | |||
702 | #[test] |
|
702 | #[test] | |
703 | fn test_block_debug() { |
|
703 | fn test_block_debug() { | |
704 | let mut block = Block::new(); |
|
704 | let mut block = Block::new(); | |
705 | block.set(1, Element::Rev(3)); |
|
705 | block.set(1, Element::Rev(3)); | |
706 | block.set(10, Element::Block(0)); |
|
706 | block.set(10, Element::Block(0)); | |
707 | assert_eq!(format!("{:?}", block), "{1: Rev(3), 10: Block(0)}"); |
|
707 | assert_eq!(format!("{:?}", block), "{1: Rev(3), 10: Block(0)}"); | |
708 | } |
|
708 | } | |
709 |
|
709 | |||
710 | #[test] |
|
710 | #[test] | |
711 | fn test_block_macro() { |
|
711 | fn test_block_macro() { | |
712 | let block = block! {5: Block(2)}; |
|
712 | let block = block! {5: Block(2)}; | |
713 | assert_eq!(format!("{:?}", block), "{5: Block(2)}"); |
|
713 | assert_eq!(format!("{:?}", block), "{5: Block(2)}"); | |
714 |
|
714 | |||
715 | let block = block! {13: Rev(15), 5: Block(2)}; |
|
715 | let block = block! {13: Rev(15), 5: Block(2)}; | |
716 | assert_eq!(format!("{:?}", block), "{5: Block(2), 13: Rev(15)}"); |
|
716 | assert_eq!(format!("{:?}", block), "{5: Block(2), 13: Rev(15)}"); | |
717 | } |
|
717 | } | |
718 |
|
718 | |||
719 | #[test] |
|
719 | #[test] | |
720 | fn test_raw_block() { |
|
720 | fn test_raw_block() { | |
721 | let mut raw = [255u8; 64]; |
|
721 | let mut raw = [255u8; 64]; | |
722 |
|
722 | |||
723 | let mut counter = 0; |
|
723 | let mut counter = 0; | |
724 | for val in [0_i32, 15, -2, -1, -3].iter() { |
|
724 | for val in [0_i32, 15, -2, -1, -3].iter() { | |
725 | for byte in val.to_be_bytes().iter() { |
|
725 | for byte in val.to_be_bytes().iter() { | |
726 | raw[counter] = *byte; |
|
726 | raw[counter] = *byte; | |
727 | counter += 1; |
|
727 | counter += 1; | |
728 | } |
|
728 | } | |
729 | } |
|
729 | } | |
730 | let (block, _) = Block::from_bytes(&raw).unwrap(); |
|
730 | let (block, _) = Block::from_bytes(&raw).unwrap(); | |
731 | assert_eq!(block.get(0), Element::Block(0)); |
|
731 | assert_eq!(block.get(0), Element::Block(0)); | |
732 | assert_eq!(block.get(1), Element::Block(15)); |
|
732 | assert_eq!(block.get(1), Element::Block(15)); | |
733 | assert_eq!(block.get(3), Element::None); |
|
733 | assert_eq!(block.get(3), Element::None); | |
734 | assert_eq!(block.get(2), Element::Rev(0)); |
|
734 | assert_eq!(block.get(2), Element::Rev(0)); | |
735 | assert_eq!(block.get(4), Element::Rev(1)); |
|
735 | assert_eq!(block.get(4), Element::Rev(1)); | |
736 | } |
|
736 | } | |
737 |
|
737 | |||
738 | type TestIndex = HashMap<Revision, Node>; |
|
738 | type TestIndex = HashMap<Revision, Node>; | |
739 |
|
739 | |||
740 | impl RevlogIndex for TestIndex { |
|
740 | impl RevlogIndex for TestIndex { | |
741 | fn node(&self, rev: Revision) -> Option<&Node> { |
|
741 | fn node(&self, rev: Revision) -> Option<&Node> { | |
742 | self.get(&rev) |
|
742 | self.get(&rev) | |
743 | } |
|
743 | } | |
744 |
|
744 | |||
745 | fn len(&self) -> usize { |
|
745 | fn len(&self) -> usize { | |
746 | self.len() |
|
746 | self.len() | |
747 | } |
|
747 | } | |
748 | } |
|
748 | } | |
749 |
|
749 | |||
750 | /// Pad hexadecimal Node prefix with zeros on the right |
|
750 | /// Pad hexadecimal Node prefix with zeros on the right | |
751 | /// |
|
751 | /// | |
752 | /// This avoids having to repeatedly write very long hexadecimal |
|
752 | /// This avoids having to repeatedly write very long hexadecimal | |
753 | /// strings for test data, and brings actual hash size independency. |
|
753 | /// strings for test data, and brings actual hash size independency. | |
754 | #[cfg(test)] |
|
754 | #[cfg(test)] | |
755 | fn pad_node(hex: &str) -> Node { |
|
755 | fn pad_node(hex: &str) -> Node { | |
756 | Node::from_hex(&hex_pad_right(hex)).unwrap() |
|
756 | Node::from_hex(&hex_pad_right(hex)).unwrap() | |
757 | } |
|
757 | } | |
758 |
|
758 | |||
759 | /// Pad hexadecimal Node prefix with zeros on the right, then insert |
|
759 | /// Pad hexadecimal Node prefix with zeros on the right, then insert | |
760 | fn pad_insert(idx: &mut TestIndex, rev: Revision, hex: &str) { |
|
760 | fn pad_insert(idx: &mut TestIndex, rev: Revision, hex: &str) { | |
761 | idx.insert(rev, pad_node(hex)); |
|
761 | idx.insert(rev, pad_node(hex)); | |
762 | } |
|
762 | } | |
763 |
|
763 | |||
764 | fn sample_nodetree() -> NodeTree { |
|
764 | fn sample_nodetree() -> NodeTree { | |
765 | NodeTree::from(vec![ |
|
765 | NodeTree::from(vec![ | |
766 | block![0: Rev(9)], |
|
766 | block![0: Rev(9)], | |
767 | block![0: Rev(0), 1: Rev(9)], |
|
767 | block![0: Rev(0), 1: Rev(9)], | |
768 | block![0: Block(1), 1:Rev(1)], |
|
768 | block![0: Block(1), 1:Rev(1)], | |
769 | ]) |
|
769 | ]) | |
770 | } |
|
770 | } | |
771 |
|
771 | |||
772 | fn hex(s: &str) -> NodePrefix { |
|
772 | fn hex(s: &str) -> NodePrefix { | |
773 | NodePrefix::from_hex(s).unwrap() |
|
773 | NodePrefix::from_hex(s).unwrap() | |
774 | } |
|
774 | } | |
775 |
|
775 | |||
776 | #[test] |
|
776 | #[test] | |
777 | fn test_nt_debug() { |
|
777 | fn test_nt_debug() { | |
778 | let nt = sample_nodetree(); |
|
778 | let nt = sample_nodetree(); | |
779 | assert_eq!( |
|
779 | assert_eq!( | |
780 | format!("{:?}", nt), |
|
780 | format!("{:?}", nt), | |
781 | "readonly: \ |
|
781 | "readonly: \ | |
782 | [{0: Rev(9)}, {0: Rev(0), 1: Rev(9)}, {0: Block(1), 1: Rev(1)}], \ |
|
782 | [{0: Rev(9)}, {0: Rev(0), 1: Rev(9)}, {0: Block(1), 1: Rev(1)}], \ | |
783 | growable: [], \ |
|
783 | growable: [], \ | |
784 | root: {0: Block(1), 1: Rev(1)}", |
|
784 | root: {0: Block(1), 1: Rev(1)}", | |
785 | ); |
|
785 | ); | |
786 | } |
|
786 | } | |
787 |
|
787 | |||
788 | #[test] |
|
788 | #[test] | |
789 | fn test_immutable_find_simplest() -> Result<(), NodeMapError> { |
|
789 | fn test_immutable_find_simplest() -> Result<(), NodeMapError> { | |
790 | let mut idx: TestIndex = HashMap::new(); |
|
790 | let mut idx: TestIndex = HashMap::new(); | |
791 | pad_insert(&mut idx, 1, "1234deadcafe"); |
|
791 | pad_insert(&mut idx, 1, "1234deadcafe"); | |
792 |
|
792 | |||
793 | let nt = NodeTree::from(vec![block! {1: Rev(1)}]); |
|
793 | let nt = NodeTree::from(vec![block! {1: Rev(1)}]); | |
794 | assert_eq!(nt.find_bin(&idx, hex("1"))?, Some(1)); |
|
794 | assert_eq!(nt.find_bin(&idx, hex("1"))?, Some(1)); | |
795 | assert_eq!(nt.find_bin(&idx, hex("12"))?, Some(1)); |
|
795 | assert_eq!(nt.find_bin(&idx, hex("12"))?, Some(1)); | |
796 | assert_eq!(nt.find_bin(&idx, hex("1234de"))?, Some(1)); |
|
796 | assert_eq!(nt.find_bin(&idx, hex("1234de"))?, Some(1)); | |
797 | assert_eq!(nt.find_bin(&idx, hex("1a"))?, None); |
|
797 | assert_eq!(nt.find_bin(&idx, hex("1a"))?, None); | |
798 | assert_eq!(nt.find_bin(&idx, hex("ab"))?, None); |
|
798 | assert_eq!(nt.find_bin(&idx, hex("ab"))?, None); | |
799 |
|
799 | |||
800 | // and with full binary Nodes |
|
800 | // and with full binary Nodes | |
801 | assert_eq!(nt.find_node(&idx, idx.get(&1).unwrap())?, Some(1)); |
|
801 | assert_eq!(nt.find_node(&idx, idx.get(&1).unwrap())?, Some(1)); | |
802 | let unknown = Node::from_hex(&hex_pad_right("3d")).unwrap(); |
|
802 | let unknown = Node::from_hex(&hex_pad_right("3d")).unwrap(); | |
803 | assert_eq!(nt.find_node(&idx, &unknown)?, None); |
|
803 | assert_eq!(nt.find_node(&idx, &unknown)?, None); | |
804 | Ok(()) |
|
804 | Ok(()) | |
805 | } |
|
805 | } | |
806 |
|
806 | |||
807 | #[test] |
|
807 | #[test] | |
808 | fn test_immutable_find_one_jump() { |
|
808 | fn test_immutable_find_one_jump() { | |
809 | let mut idx = TestIndex::new(); |
|
809 | let mut idx = TestIndex::new(); | |
810 | pad_insert(&mut idx, 9, "012"); |
|
810 | pad_insert(&mut idx, 9, "012"); | |
811 | pad_insert(&mut idx, 0, "00a"); |
|
811 | pad_insert(&mut idx, 0, "00a"); | |
812 |
|
812 | |||
813 | let nt = sample_nodetree(); |
|
813 | let nt = sample_nodetree(); | |
814 |
|
814 | |||
815 | assert_eq!(nt.find_bin(&idx, hex("0")), Err(MultipleResults)); |
|
815 | assert_eq!(nt.find_bin(&idx, hex("0")), Err(MultipleResults)); | |
816 | assert_eq!(nt.find_bin(&idx, hex("01")), Ok(Some(9))); |
|
816 | assert_eq!(nt.find_bin(&idx, hex("01")), Ok(Some(9))); | |
817 | assert_eq!(nt.find_bin(&idx, hex("00")), Err(MultipleResults)); |
|
817 | assert_eq!(nt.find_bin(&idx, hex("00")), Err(MultipleResults)); | |
818 | assert_eq!(nt.find_bin(&idx, hex("00a")), Ok(Some(0))); |
|
818 | assert_eq!(nt.find_bin(&idx, hex("00a")), Ok(Some(0))); | |
819 | assert_eq!(nt.unique_prefix_len_bin(&idx, hex("00a")), Ok(Some(3))); |
|
819 | assert_eq!(nt.unique_prefix_len_bin(&idx, hex("00a")), Ok(Some(3))); | |
820 | assert_eq!(nt.find_bin(&idx, hex("000")), Ok(Some(NULL_REVISION))); |
|
820 | assert_eq!(nt.find_bin(&idx, hex("000")), Ok(Some(NULL_REVISION))); | |
821 | } |
|
821 | } | |
822 |
|
822 | |||
823 | #[test] |
|
823 | #[test] | |
824 | fn test_mutated_find() -> Result<(), NodeMapError> { |
|
824 | fn test_mutated_find() -> Result<(), NodeMapError> { | |
825 | let mut idx = TestIndex::new(); |
|
825 | let mut idx = TestIndex::new(); | |
826 | pad_insert(&mut idx, 9, "012"); |
|
826 | pad_insert(&mut idx, 9, "012"); | |
827 | pad_insert(&mut idx, 0, "00a"); |
|
827 | pad_insert(&mut idx, 0, "00a"); | |
828 | pad_insert(&mut idx, 2, "cafe"); |
|
828 | pad_insert(&mut idx, 2, "cafe"); | |
829 | pad_insert(&mut idx, 3, "15"); |
|
829 | pad_insert(&mut idx, 3, "15"); | |
830 | pad_insert(&mut idx, 1, "10"); |
|
830 | pad_insert(&mut idx, 1, "10"); | |
831 |
|
831 | |||
832 | let nt = NodeTree { |
|
832 | let nt = NodeTree { | |
833 | readonly: sample_nodetree().readonly, |
|
833 | readonly: sample_nodetree().readonly, | |
834 | growable: vec![block![0: Rev(1), 5: Rev(3)]], |
|
834 | growable: vec![block![0: Rev(1), 5: Rev(3)]], | |
835 | root: block![0: Block(1), 1:Block(3), 12: Rev(2)], |
|
835 | root: block![0: Block(1), 1:Block(3), 12: Rev(2)], | |
836 | masked_inner_blocks: 1, |
|
836 | masked_inner_blocks: 1, | |
837 | }; |
|
837 | }; | |
838 | assert_eq!(nt.find_bin(&idx, hex("10"))?, Some(1)); |
|
838 | assert_eq!(nt.find_bin(&idx, hex("10"))?, Some(1)); | |
839 | assert_eq!(nt.find_bin(&idx, hex("c"))?, Some(2)); |
|
839 | assert_eq!(nt.find_bin(&idx, hex("c"))?, Some(2)); | |
840 | assert_eq!(nt.unique_prefix_len_bin(&idx, hex("c"))?, Some(1)); |
|
840 | assert_eq!(nt.unique_prefix_len_bin(&idx, hex("c"))?, Some(1)); | |
841 | assert_eq!(nt.find_bin(&idx, hex("00")), Err(MultipleResults)); |
|
841 | assert_eq!(nt.find_bin(&idx, hex("00")), Err(MultipleResults)); | |
842 | assert_eq!(nt.find_bin(&idx, hex("000"))?, Some(NULL_REVISION)); |
|
842 | assert_eq!(nt.find_bin(&idx, hex("000"))?, Some(NULL_REVISION)); | |
843 | assert_eq!(nt.unique_prefix_len_bin(&idx, hex("000"))?, Some(3)); |
|
843 | assert_eq!(nt.unique_prefix_len_bin(&idx, hex("000"))?, Some(3)); | |
844 | assert_eq!(nt.find_bin(&idx, hex("01"))?, Some(9)); |
|
844 | assert_eq!(nt.find_bin(&idx, hex("01"))?, Some(9)); | |
845 | assert_eq!(nt.masked_readonly_blocks(), 2); |
|
845 | assert_eq!(nt.masked_readonly_blocks(), 2); | |
846 | Ok(()) |
|
846 | Ok(()) | |
847 | } |
|
847 | } | |
848 |
|
848 | |||
849 | struct TestNtIndex { |
|
849 | struct TestNtIndex { | |
850 | index: TestIndex, |
|
850 | index: TestIndex, | |
851 | nt: NodeTree, |
|
851 | nt: NodeTree, | |
852 | } |
|
852 | } | |
853 |
|
853 | |||
854 | impl TestNtIndex { |
|
854 | impl TestNtIndex { | |
855 | fn new() -> Self { |
|
855 | fn new() -> Self { | |
856 | TestNtIndex { |
|
856 | TestNtIndex { | |
857 | index: HashMap::new(), |
|
857 | index: HashMap::new(), | |
858 | nt: NodeTree::default(), |
|
858 | nt: NodeTree::default(), | |
859 | } |
|
859 | } | |
860 | } |
|
860 | } | |
861 |
|
861 | |||
862 | fn insert( |
|
862 | fn insert( | |
863 | &mut self, |
|
863 | &mut self, | |
864 | rev: Revision, |
|
864 | rev: Revision, | |
865 | hex: &str, |
|
865 | hex: &str, | |
866 | ) -> Result<(), NodeMapError> { |
|
866 | ) -> Result<(), NodeMapError> { | |
867 | let node = pad_node(hex); |
|
867 | let node = pad_node(hex); | |
868 | self.index.insert(rev, node.clone()); |
|
868 | self.index.insert(rev, node.clone()); | |
869 | self.nt.insert(&self.index, &node, rev)?; |
|
869 | self.nt.insert(&self.index, &node, rev)?; | |
870 | Ok(()) |
|
870 | Ok(()) | |
871 | } |
|
871 | } | |
872 |
|
872 | |||
873 | fn find_hex( |
|
873 | fn find_hex( | |
874 | &self, |
|
874 | &self, | |
875 | prefix: &str, |
|
875 | prefix: &str, | |
876 | ) -> Result<Option<Revision>, NodeMapError> { |
|
876 | ) -> Result<Option<Revision>, NodeMapError> { | |
877 | self.nt.find_bin(&self.index, hex(prefix)) |
|
877 | self.nt.find_bin(&self.index, hex(prefix)) | |
878 | } |
|
878 | } | |
879 |
|
879 | |||
880 | fn unique_prefix_len_hex( |
|
880 | fn unique_prefix_len_hex( | |
881 | &self, |
|
881 | &self, | |
882 | prefix: &str, |
|
882 | prefix: &str, | |
883 | ) -> Result<Option<usize>, NodeMapError> { |
|
883 | ) -> Result<Option<usize>, NodeMapError> { | |
884 | self.nt.unique_prefix_len_bin(&self.index, hex(prefix)) |
|
884 | self.nt.unique_prefix_len_bin(&self.index, hex(prefix)) | |
885 | } |
|
885 | } | |
886 |
|
886 | |||
887 | /// Drain `added` and restart a new one |
|
887 | /// Drain `added` and restart a new one | |
888 | fn commit(self) -> Self { |
|
888 | fn commit(self) -> Self { | |
889 | let mut as_vec: Vec<Block> = |
|
889 | let mut as_vec: Vec<Block> = | |
890 | self.nt.readonly.iter().map(|block| block.clone()).collect(); |
|
890 | self.nt.readonly.iter().map(|block| block.clone()).collect(); | |
891 | as_vec.extend(self.nt.growable); |
|
891 | as_vec.extend(self.nt.growable); | |
892 | as_vec.push(self.nt.root); |
|
892 | as_vec.push(self.nt.root); | |
893 |
|
893 | |||
894 | Self { |
|
894 | Self { | |
895 | index: self.index, |
|
895 | index: self.index, | |
896 | nt: NodeTree::from(as_vec).into(), |
|
896 | nt: NodeTree::from(as_vec).into(), | |
897 | } |
|
897 | } | |
898 | } |
|
898 | } | |
899 | } |
|
899 | } | |
900 |
|
900 | |||
901 | #[test] |
|
901 | #[test] | |
902 | fn test_insert_full_mutable() -> Result<(), NodeMapError> { |
|
902 | fn test_insert_full_mutable() -> Result<(), NodeMapError> { | |
903 | let mut idx = TestNtIndex::new(); |
|
903 | let mut idx = TestNtIndex::new(); | |
904 | idx.insert(0, "1234")?; |
|
904 | idx.insert(0, "1234")?; | |
905 | assert_eq!(idx.find_hex("1")?, Some(0)); |
|
905 | assert_eq!(idx.find_hex("1")?, Some(0)); | |
906 | assert_eq!(idx.find_hex("12")?, Some(0)); |
|
906 | assert_eq!(idx.find_hex("12")?, Some(0)); | |
907 |
|
907 | |||
908 | // let's trigger a simple split |
|
908 | // let's trigger a simple split | |
909 | idx.insert(1, "1a34")?; |
|
909 | idx.insert(1, "1a34")?; | |
910 | assert_eq!(idx.nt.growable.len(), 1); |
|
910 | assert_eq!(idx.nt.growable.len(), 1); | |
911 | assert_eq!(idx.find_hex("12")?, Some(0)); |
|
911 | assert_eq!(idx.find_hex("12")?, Some(0)); | |
912 | assert_eq!(idx.find_hex("1a")?, Some(1)); |
|
912 | assert_eq!(idx.find_hex("1a")?, Some(1)); | |
913 |
|
913 | |||
914 | // reinserting is a no_op |
|
914 | // reinserting is a no_op | |
915 | idx.insert(1, "1a34")?; |
|
915 | idx.insert(1, "1a34")?; | |
916 | assert_eq!(idx.nt.growable.len(), 1); |
|
916 | assert_eq!(idx.nt.growable.len(), 1); | |
917 | assert_eq!(idx.find_hex("12")?, Some(0)); |
|
917 | assert_eq!(idx.find_hex("12")?, Some(0)); | |
918 | assert_eq!(idx.find_hex("1a")?, Some(1)); |
|
918 | assert_eq!(idx.find_hex("1a")?, Some(1)); | |
919 |
|
919 | |||
920 | idx.insert(2, "1a01")?; |
|
920 | idx.insert(2, "1a01")?; | |
921 | assert_eq!(idx.nt.growable.len(), 2); |
|
921 | assert_eq!(idx.nt.growable.len(), 2); | |
922 | assert_eq!(idx.find_hex("1a"), Err(NodeMapError::MultipleResults)); |
|
922 | assert_eq!(idx.find_hex("1a"), Err(NodeMapError::MultipleResults)); | |
923 | assert_eq!(idx.find_hex("12")?, Some(0)); |
|
923 | assert_eq!(idx.find_hex("12")?, Some(0)); | |
924 | assert_eq!(idx.find_hex("1a3")?, Some(1)); |
|
924 | assert_eq!(idx.find_hex("1a3")?, Some(1)); | |
925 | assert_eq!(idx.find_hex("1a0")?, Some(2)); |
|
925 | assert_eq!(idx.find_hex("1a0")?, Some(2)); | |
926 | assert_eq!(idx.find_hex("1a12")?, None); |
|
926 | assert_eq!(idx.find_hex("1a12")?, None); | |
927 |
|
927 | |||
928 | // now let's make it split and create more than one additional block |
|
928 | // now let's make it split and create more than one additional block | |
929 | idx.insert(3, "1a345")?; |
|
929 | idx.insert(3, "1a345")?; | |
930 | assert_eq!(idx.nt.growable.len(), 4); |
|
930 | assert_eq!(idx.nt.growable.len(), 4); | |
931 | assert_eq!(idx.find_hex("1a340")?, Some(1)); |
|
931 | assert_eq!(idx.find_hex("1a340")?, Some(1)); | |
932 | assert_eq!(idx.find_hex("1a345")?, Some(3)); |
|
932 | assert_eq!(idx.find_hex("1a345")?, Some(3)); | |
933 | assert_eq!(idx.find_hex("1a341")?, None); |
|
933 | assert_eq!(idx.find_hex("1a341")?, None); | |
934 |
|
934 | |||
935 | // there's no readonly block to mask |
|
935 | // there's no readonly block to mask | |
936 | assert_eq!(idx.nt.masked_readonly_blocks(), 0); |
|
936 | assert_eq!(idx.nt.masked_readonly_blocks(), 0); | |
937 | Ok(()) |
|
937 | Ok(()) | |
938 | } |
|
938 | } | |
939 |
|
939 | |||
940 | #[test] |
|
940 | #[test] | |
941 | fn test_unique_prefix_len_zero_prefix() { |
|
941 | fn test_unique_prefix_len_zero_prefix() { | |
942 | let mut idx = TestNtIndex::new(); |
|
942 | let mut idx = TestNtIndex::new(); | |
943 | idx.insert(0, "00000abcd").unwrap(); |
|
943 | idx.insert(0, "00000abcd").unwrap(); | |
944 |
|
944 | |||
945 | assert_eq!(idx.find_hex("000"), Err(NodeMapError::MultipleResults)); |
|
945 | assert_eq!(idx.find_hex("000"), Err(NodeMapError::MultipleResults)); | |
946 | // in the nodetree proper, this will be found at the first nybble |
|
946 | // in the nodetree proper, this will be found at the first nybble | |
947 | // yet the correct answer for unique_prefix_len is not 1, nor 1+1, |
|
947 | // yet the correct answer for unique_prefix_len is not 1, nor 1+1, | |
948 | // but the first difference with `NULL_NODE` |
|
948 | // but the first difference with `NULL_NODE` | |
949 | assert_eq!(idx.unique_prefix_len_hex("00000a"), Ok(Some(6))); |
|
949 | assert_eq!(idx.unique_prefix_len_hex("00000a"), Ok(Some(6))); | |
950 | assert_eq!(idx.unique_prefix_len_hex("00000ab"), Ok(Some(6))); |
|
950 | assert_eq!(idx.unique_prefix_len_hex("00000ab"), Ok(Some(6))); | |
951 |
|
951 | |||
952 | // same with odd result |
|
952 | // same with odd result | |
953 | idx.insert(1, "00123").unwrap(); |
|
953 | idx.insert(1, "00123").unwrap(); | |
954 | assert_eq!(idx.unique_prefix_len_hex("001"), Ok(Some(3))); |
|
954 | assert_eq!(idx.unique_prefix_len_hex("001"), Ok(Some(3))); | |
955 | assert_eq!(idx.unique_prefix_len_hex("0012"), Ok(Some(3))); |
|
955 | assert_eq!(idx.unique_prefix_len_hex("0012"), Ok(Some(3))); | |
956 |
|
956 | |||
957 | // these are unchanged of course |
|
957 | // these are unchanged of course | |
958 | assert_eq!(idx.unique_prefix_len_hex("00000a"), Ok(Some(6))); |
|
958 | assert_eq!(idx.unique_prefix_len_hex("00000a"), Ok(Some(6))); | |
959 | assert_eq!(idx.unique_prefix_len_hex("00000ab"), Ok(Some(6))); |
|
959 | assert_eq!(idx.unique_prefix_len_hex("00000ab"), Ok(Some(6))); | |
960 | } |
|
960 | } | |
961 |
|
961 | |||
962 | #[test] |
|
962 | #[test] | |
963 | fn test_insert_extreme_splitting() -> Result<(), NodeMapError> { |
|
963 | fn test_insert_extreme_splitting() -> Result<(), NodeMapError> { | |
964 | // check that the splitting loop is long enough |
|
964 | // check that the splitting loop is long enough | |
965 | let mut nt_idx = TestNtIndex::new(); |
|
965 | let mut nt_idx = TestNtIndex::new(); | |
966 | let nt = &mut nt_idx.nt; |
|
966 | let nt = &mut nt_idx.nt; | |
967 | let idx = &mut nt_idx.index; |
|
967 | let idx = &mut nt_idx.index; | |
968 |
|
968 | |||
969 | let node0_hex = hex_pad_right("444444"); |
|
969 | let node0_hex = hex_pad_right("444444"); | |
970 | let mut node1_hex = hex_pad_right("444444").clone(); |
|
970 | let mut node1_hex = hex_pad_right("444444").clone(); | |
971 | node1_hex.pop(); |
|
971 | node1_hex.pop(); | |
972 | node1_hex.push('5'); |
|
972 | node1_hex.push('5'); | |
973 | let node0 = Node::from_hex(&node0_hex).unwrap(); |
|
973 | let node0 = Node::from_hex(&node0_hex).unwrap(); | |
974 | let node1 = Node::from_hex(&node1_hex).unwrap(); |
|
974 | let node1 = Node::from_hex(&node1_hex).unwrap(); | |
975 |
|
975 | |||
976 | idx.insert(0, node0.clone()); |
|
976 | idx.insert(0, node0.clone()); | |
977 | nt.insert(idx, &node0, 0)?; |
|
977 | nt.insert(idx, &node0, 0)?; | |
978 | idx.insert(1, node1.clone()); |
|
978 | idx.insert(1, node1.clone()); | |
979 | nt.insert(idx, &node1, 1)?; |
|
979 | nt.insert(idx, &node1, 1)?; | |
980 |
|
980 | |||
981 | assert_eq!(nt.find_bin(idx, (&node0).into())?, Some(0)); |
|
981 | assert_eq!(nt.find_bin(idx, (&node0).into())?, Some(0)); | |
982 | assert_eq!(nt.find_bin(idx, (&node1).into())?, Some(1)); |
|
982 | assert_eq!(nt.find_bin(idx, (&node1).into())?, Some(1)); | |
983 | Ok(()) |
|
983 | Ok(()) | |
984 | } |
|
984 | } | |
985 |
|
985 | |||
986 | #[test] |
|
986 | #[test] | |
987 | fn test_insert_partly_immutable() -> Result<(), NodeMapError> { |
|
987 | fn test_insert_partly_immutable() -> Result<(), NodeMapError> { | |
988 | let mut idx = TestNtIndex::new(); |
|
988 | let mut idx = TestNtIndex::new(); | |
989 | idx.insert(0, "1234")?; |
|
989 | idx.insert(0, "1234")?; | |
990 | idx.insert(1, "1235")?; |
|
990 | idx.insert(1, "1235")?; | |
991 | idx.insert(2, "131")?; |
|
991 | idx.insert(2, "131")?; | |
992 | idx.insert(3, "cafe")?; |
|
992 | idx.insert(3, "cafe")?; | |
993 | let mut idx = idx.commit(); |
|
993 | let mut idx = idx.commit(); | |
994 | assert_eq!(idx.find_hex("1234")?, Some(0)); |
|
994 | assert_eq!(idx.find_hex("1234")?, Some(0)); | |
995 | assert_eq!(idx.find_hex("1235")?, Some(1)); |
|
995 | assert_eq!(idx.find_hex("1235")?, Some(1)); | |
996 | assert_eq!(idx.find_hex("131")?, Some(2)); |
|
996 | assert_eq!(idx.find_hex("131")?, Some(2)); | |
997 | assert_eq!(idx.find_hex("cafe")?, Some(3)); |
|
997 | assert_eq!(idx.find_hex("cafe")?, Some(3)); | |
998 | // we did not add anything since init from readonly |
|
998 | // we did not add anything since init from readonly | |
999 | assert_eq!(idx.nt.masked_readonly_blocks(), 0); |
|
999 | assert_eq!(idx.nt.masked_readonly_blocks(), 0); | |
1000 |
|
1000 | |||
1001 | idx.insert(4, "123A")?; |
|
1001 | idx.insert(4, "123A")?; | |
1002 | assert_eq!(idx.find_hex("1234")?, Some(0)); |
|
1002 | assert_eq!(idx.find_hex("1234")?, Some(0)); | |
1003 | assert_eq!(idx.find_hex("1235")?, Some(1)); |
|
1003 | assert_eq!(idx.find_hex("1235")?, Some(1)); | |
1004 | assert_eq!(idx.find_hex("131")?, Some(2)); |
|
1004 | assert_eq!(idx.find_hex("131")?, Some(2)); | |
1005 | assert_eq!(idx.find_hex("cafe")?, Some(3)); |
|
1005 | assert_eq!(idx.find_hex("cafe")?, Some(3)); | |
1006 | assert_eq!(idx.find_hex("123A")?, Some(4)); |
|
1006 | assert_eq!(idx.find_hex("123A")?, Some(4)); | |
1007 | // we masked blocks for all prefixes of "123", including the root |
|
1007 | // we masked blocks for all prefixes of "123", including the root | |
1008 | assert_eq!(idx.nt.masked_readonly_blocks(), 4); |
|
1008 | assert_eq!(idx.nt.masked_readonly_blocks(), 4); | |
1009 |
|
1009 | |||
1010 | eprintln!("{:?}", idx.nt); |
|
1010 | eprintln!("{:?}", idx.nt); | |
1011 | idx.insert(5, "c0")?; |
|
1011 | idx.insert(5, "c0")?; | |
1012 | assert_eq!(idx.find_hex("cafe")?, Some(3)); |
|
1012 | assert_eq!(idx.find_hex("cafe")?, Some(3)); | |
1013 | assert_eq!(idx.find_hex("c0")?, Some(5)); |
|
1013 | assert_eq!(idx.find_hex("c0")?, Some(5)); | |
1014 | assert_eq!(idx.find_hex("c1")?, None); |
|
1014 | assert_eq!(idx.find_hex("c1")?, None); | |
1015 | assert_eq!(idx.find_hex("1234")?, Some(0)); |
|
1015 | assert_eq!(idx.find_hex("1234")?, Some(0)); | |
1016 | // inserting "c0" is just splitting the 'c' slot of the mutable root, |
|
1016 | // inserting "c0" is just splitting the 'c' slot of the mutable root, | |
1017 | // it doesn't mask anything |
|
1017 | // it doesn't mask anything | |
1018 | assert_eq!(idx.nt.masked_readonly_blocks(), 4); |
|
1018 | assert_eq!(idx.nt.masked_readonly_blocks(), 4); | |
1019 |
|
1019 | |||
1020 | Ok(()) |
|
1020 | Ok(()) | |
1021 | } |
|
1021 | } | |
1022 |
|
1022 | |||
1023 | #[test] |
|
1023 | #[test] | |
1024 | fn test_invalidate_all() -> Result<(), NodeMapError> { |
|
1024 | fn test_invalidate_all() -> Result<(), NodeMapError> { | |
1025 | let mut idx = TestNtIndex::new(); |
|
1025 | let mut idx = TestNtIndex::new(); | |
1026 | idx.insert(0, "1234")?; |
|
1026 | idx.insert(0, "1234")?; | |
1027 | idx.insert(1, "1235")?; |
|
1027 | idx.insert(1, "1235")?; | |
1028 | idx.insert(2, "131")?; |
|
1028 | idx.insert(2, "131")?; | |
1029 | idx.insert(3, "cafe")?; |
|
1029 | idx.insert(3, "cafe")?; | |
1030 | let mut idx = idx.commit(); |
|
1030 | let mut idx = idx.commit(); | |
1031 |
|
1031 | |||
1032 | idx.nt.invalidate_all(); |
|
1032 | idx.nt.invalidate_all(); | |
1033 |
|
1033 | |||
1034 | assert_eq!(idx.find_hex("1234")?, None); |
|
1034 | assert_eq!(idx.find_hex("1234")?, None); | |
1035 | assert_eq!(idx.find_hex("1235")?, None); |
|
1035 | assert_eq!(idx.find_hex("1235")?, None); | |
1036 | assert_eq!(idx.find_hex("131")?, None); |
|
1036 | assert_eq!(idx.find_hex("131")?, None); | |
1037 | assert_eq!(idx.find_hex("cafe")?, None); |
|
1037 | assert_eq!(idx.find_hex("cafe")?, None); | |
1038 | // all the readonly blocks have been masked, this is the |
|
1038 | // all the readonly blocks have been masked, this is the | |
1039 | // conventional expected response |
|
1039 | // conventional expected response | |
1040 | assert_eq!(idx.nt.masked_readonly_blocks(), idx.nt.readonly.len() + 1); |
|
1040 | assert_eq!(idx.nt.masked_readonly_blocks(), idx.nt.readonly.len() + 1); | |
1041 | Ok(()) |
|
1041 | Ok(()) | |
1042 | } |
|
1042 | } | |
1043 |
|
1043 | |||
1044 | #[test] |
|
1044 | #[test] | |
1045 | fn test_into_added_empty() { |
|
1045 | fn test_into_added_empty() { | |
1046 | assert!(sample_nodetree().into_readonly_and_added().1.is_empty()); |
|
1046 | assert!(sample_nodetree().into_readonly_and_added().1.is_empty()); | |
1047 | assert!(sample_nodetree() |
|
1047 | assert!(sample_nodetree() | |
1048 | .into_readonly_and_added_bytes() |
|
1048 | .into_readonly_and_added_bytes() | |
1049 | .1 |
|
1049 | .1 | |
1050 | .is_empty()); |
|
1050 | .is_empty()); | |
1051 | } |
|
1051 | } | |
1052 |
|
1052 | |||
1053 | #[test] |
|
1053 | #[test] | |
1054 | fn test_into_added_bytes() -> Result<(), NodeMapError> { |
|
1054 | fn test_into_added_bytes() -> Result<(), NodeMapError> { | |
1055 | let mut idx = TestNtIndex::new(); |
|
1055 | let mut idx = TestNtIndex::new(); | |
1056 | idx.insert(0, "1234")?; |
|
1056 | idx.insert(0, "1234")?; | |
1057 | let mut idx = idx.commit(); |
|
1057 | let mut idx = idx.commit(); | |
1058 | idx.insert(4, "cafe")?; |
|
1058 | idx.insert(4, "cafe")?; | |
1059 | let (_, bytes) = idx.nt.into_readonly_and_added_bytes(); |
|
1059 | let (_, bytes) = idx.nt.into_readonly_and_added_bytes(); | |
1060 |
|
1060 | |||
1061 | // only the root block has been changed |
|
1061 | // only the root block has been changed | |
1062 | assert_eq!(bytes.len(), size_of::<Block>()); |
|
1062 | assert_eq!(bytes.len(), size_of::<Block>()); | |
1063 | // big endian for -2 |
|
1063 | // big endian for -2 | |
1064 | assert_eq!(&bytes[4..2 * 4], [255, 255, 255, 254]); |
|
1064 | assert_eq!(&bytes[4..2 * 4], [255, 255, 255, 254]); | |
1065 | // big endian for -6 |
|
1065 | // big endian for -6 | |
1066 | assert_eq!(&bytes[12 * 4..13 * 4], [255, 255, 255, 250]); |
|
1066 | assert_eq!(&bytes[12 * 4..13 * 4], [255, 255, 255, 250]); | |
1067 | Ok(()) |
|
1067 | Ok(()) | |
1068 | } |
|
1068 | } | |
1069 | } |
|
1069 | } |
General Comments 0
You need to be logged in to leave comments.
Login now