Mercurial > public > mercurial-scm > hg-stable

annotate rust/hg-core/src/revlog/filelog.rs @ 48575:e91aa800ae5b

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
rhg: desambiguate status without decompressing filelog if possible When status is unsure based on `stat()` and the dirstate if a file is clean or modified, we need to compare it against the filelog. This comparison can skip looking at contents if the lengths differ. This changeset optimize this further to deduce what we can about the length if the filelog without decompressing it or resolving deltas. Differential Revision: https://phab.mercurial-scm.org/D11965
author Simon Sapin <simon.sapin@octobus.net>
date Fri, 07 Jan 2022 14:40:21 +0100
parents 35c47015b9b7
children bfc117647c71
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
1 use crate::errors::HgError;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
2 use crate::repo::Repo;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
3 use crate::revlog::path_encode::path_encode;
48571
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
4 use crate::revlog::revlog::RevlogEntry;
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
5 use crate::revlog::revlog::{Revlog, RevlogError};
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
6 use crate::revlog::NodePrefix;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
7 use crate::revlog::Revision;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
8 use crate::utils::files::get_path_from_bytes;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
9 use crate::utils::hg_path::HgPath;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
10 use crate::utils::SliceExt;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
11 use std::path::PathBuf;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
12
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
13 /// A specialized `Revlog` to work with file data logs.
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
14 pub struct Filelog {
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
15 /// The generic `revlog` format.
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
16 revlog: Revlog,
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
17 }
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
18
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
19 impl Filelog {
47991
001d747c2baf rust: Return HgError instead of RevlogError in revlog constructors
Simon Sapin <simon.sapin@octobus.net>
parents: 47989
diff changeset
20 pub fn open(repo: &Repo, file_path: &HgPath) -> Result<Self, HgError> {
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
21 let index_path = store_path(file_path, b".i");
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
22 let data_path = store_path(file_path, b".d");
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
23 let revlog = Revlog::open(repo, index_path, Some(&data_path))?;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
24 Ok(Self { revlog })
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
25 }
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
26
48571
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
27 /// The given node ID is that of the file as found in a filelog, not of a
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
28 /// changeset.
47997
87e3f878e65f rust: Rename get_node methods to data_for_node, get_rev to data_for_rev
Simon Sapin <simon.sapin@octobus.net>
parents: 47996
diff changeset
29 pub fn data_for_node(
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
30 &self,
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
31 file_node: impl Into<NodePrefix>,
48569
20d0d896183e rhg: Rename some revlog-related types and methods
Simon Sapin <simon.sapin@octobus.net>
parents: 48260
diff changeset
32 ) -> Result<FilelogRevisionData, RevlogError> {
47996
6f579618ea7b rust: Rename the `Revlog::get_node_rev` method to `rev_from_node`
Simon Sapin <simon.sapin@octobus.net>
parents: 47991
diff changeset
33 let file_rev = self.revlog.rev_from_node(file_node.into())?;
47997
87e3f878e65f rust: Rename get_node methods to data_for_node, get_rev to data_for_rev
Simon Sapin <simon.sapin@octobus.net>
parents: 47996
diff changeset
34 self.data_for_rev(file_rev)
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
35 }
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
36
48571
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
37 /// The given revision is that of the file as found in a filelog, not of a
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
38 /// changeset.
47997
87e3f878e65f rust: Rename get_node methods to data_for_node, get_rev to data_for_rev
Simon Sapin <simon.sapin@octobus.net>
parents: 47996
diff changeset
39 pub fn data_for_rev(
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
40 &self,
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
41 file_rev: Revision,
48569
20d0d896183e rhg: Rename some revlog-related types and methods
Simon Sapin <simon.sapin@octobus.net>
parents: 48260
diff changeset
42 ) -> Result<FilelogRevisionData, RevlogError> {
48570
f2f57724d4eb rhg: Add RevlogEntry::data that does delta resolution
Simon Sapin <simon.sapin@octobus.net>
parents: 48569
diff changeset
43 let data: Vec<u8> = self.revlog.get_rev_data(file_rev)?.into_owned();
48569
20d0d896183e rhg: Rename some revlog-related types and methods
Simon Sapin <simon.sapin@octobus.net>
parents: 48260
diff changeset
44 Ok(FilelogRevisionData(data.into()))
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
45 }
48571
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
46
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
47 /// The given node ID is that of the file as found in a filelog, not of a
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
48 /// changeset.
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
49 pub fn entry_for_node(
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
50 &self,
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
51 file_node: impl Into<NodePrefix>,
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
52 ) -> Result<FilelogEntry, RevlogError> {
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
53 let file_rev = self.revlog.rev_from_node(file_node.into())?;
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
54 self.entry_for_rev(file_rev)
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
55 }
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
56
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
57 /// The given revision is that of the file as found in a filelog, not of a
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
58 /// changeset.
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
59 pub fn entry_for_rev(
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
60 &self,
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
61 file_rev: Revision,
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
62 ) -> Result<FilelogEntry, RevlogError> {
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
63 Ok(FilelogEntry(self.revlog.get_entry(file_rev)?))
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
64 }
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
65 }
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
66
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
67 fn store_path(hg_path: &HgPath, suffix: &[u8]) -> PathBuf {
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
68 let encoded_bytes =
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
69 path_encode(&[b"data/", hg_path.as_bytes(), suffix].concat());
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
70 get_path_from_bytes(&encoded_bytes).into()
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
71 }
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
72
48571
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
73 pub struct FilelogEntry<'a>(RevlogEntry<'a>);
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
74
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
75 impl FilelogEntry<'_> {
48575
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
76 /// `self.data()` can be expensive, with decompression and delta
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
77 /// resolution.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
78 ///
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
79 /// *Without* paying this cost, based on revlog index information
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
80 /// including `RevlogEntry::uncompressed_len`:
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
81 ///
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
82 /// * Returns `true` if the length that `self.data().file_data().len()`
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
83 /// would return is definitely **not equal** to `other_len`.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
84 /// * Returns `false` if available information is inconclusive.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
85 pub fn file_data_len_not_equal_to(&self, other_len: u64) -> bool {
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
86 // Relevant code that implement this behavior in Python code:
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
87 // basefilectx.cmp, filelog.size, storageutil.filerevisioncopied,
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
88 // revlog.size, revlog.rawsize
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
89
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
90 // Let’s call `file_data_len` what would be returned by
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
91 // `self.data().file_data().len()`.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
92
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
93 if self.0.is_cencored() {
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
94 let file_data_len = 0;
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
95 return other_len != file_data_len;
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
96 }
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
97
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
98 if self.0.has_length_affecting_flag_processor() {
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
99 // We can’t conclude anything about `file_data_len`.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
100 return false;
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
101 }
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
102
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
103 // Revlog revisions (usually) have metadata for the size of
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
104 // their data after decompression and delta resolution
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
105 // as would be returned by `Revlog::get_rev_data`.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
106 //
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
107 // For filelogs this is the file’s contents preceded by an optional
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
108 // metadata block.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
109 let uncompressed_len = if let Some(l) = self.0.uncompressed_len() {
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
110 l as u64
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
111 } else {
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
112 // The field was set to -1, the actual uncompressed len is unknown.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
113 // We need to decompress to say more.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
114 return false;
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
115 };
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
116 // `uncompressed_len = file_data_len + optional_metadata_len`,
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
117 // so `file_data_len <= uncompressed_len`.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
118 if uncompressed_len < other_len {
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
119 // Transitively, `file_data_len < other_len`.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
120 // So `other_len != file_data_len` definitely.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
121 return true;
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
122 }
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
123
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
124 if uncompressed_len == other_len + 4 {
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
125 // It’s possible that `file_data_len == other_len` with an empty
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
126 // metadata block (2 start marker bytes + 2 end marker bytes).
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
127 // This happens when there wouldn’t otherwise be metadata, but
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
128 // the first 2 bytes of file data happen to match a start marker
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
129 // and would be ambiguous.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
130 return false;
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
131 }
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
132
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
133 if !self.0.has_p1() {
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
134 // There may or may not be copy metadata, so we can’t deduce more
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
135 // about `file_data_len` without computing file data.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
136 return false;
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
137 }
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
138
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
139 // Filelog ancestry is not meaningful in the way changelog ancestry is.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
140 // It only provides hints to delta generation.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
141 // p1 and p2 are set to null when making a copy or rename since
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
142 // contents are likely unrelatedto what might have previously existed
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
143 // at the destination path.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
144 //
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
145 // Conversely, since here p1 is non-null, there is no copy metadata.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
146 // Note that this reasoning may be invalidated in the presence of
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
147 // merges made by some previous versions of Mercurial that
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
148 // swapped p1 and p2. See <https://bz.mercurial-scm.org/show_bug.cgi?id=6528>
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
149 // and `tests/test-issue6528.t`.
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
150 //
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
151 // Since copy metadata is currently the only kind of metadata
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
152 // kept in revlog data of filelogs,
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
153 // this `FilelogEntry` does not have such metadata:
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
154 let file_data_len = uncompressed_len;
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
155
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
156 return file_data_len != other_len;
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
157 }
e91aa800ae5b rhg: desambiguate status without decompressing filelog if possible
Simon Sapin <simon.sapin@octobus.net>
parents: 48571
diff changeset
158
48571
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
159 pub fn data(&self) -> Result<FilelogRevisionData, HgError> {
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
160 Ok(FilelogRevisionData(self.0.data()?.into_owned()))
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
161 }
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
162 }
35c47015b9b7 rhg: Expose FilelogEntry that wraps RevlogEntry
Simon Sapin <simon.sapin@octobus.net>
parents: 48570
diff changeset
163
48569
20d0d896183e rhg: Rename some revlog-related types and methods
Simon Sapin <simon.sapin@octobus.net>
parents: 48260
diff changeset
164 /// The data for one revision in a filelog, uncompressed and delta-resolved.
20d0d896183e rhg: Rename some revlog-related types and methods
Simon Sapin <simon.sapin@octobus.net>
parents: 48260
diff changeset
165 pub struct FilelogRevisionData(Vec<u8>);
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
166
48569
20d0d896183e rhg: Rename some revlog-related types and methods
Simon Sapin <simon.sapin@octobus.net>
parents: 48260
diff changeset
167 impl FilelogRevisionData {
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
168 /// Split into metadata and data
48260
e9faae0f445c rhg: simplify split_metadata
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 48249
diff changeset
169 pub fn split(&self) -> Result<(Option<&[u8]>, &[u8]), HgError> {
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
170 const DELIMITER: &[u8; 2] = &[b'\x01', b'\n'];
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
171
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
172 if let Some(rest) = self.0.drop_prefix(DELIMITER) {
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
173 if let Some((metadata, data)) = rest.split_2_by_slice(DELIMITER) {
48260
e9faae0f445c rhg: simplify split_metadata
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 48249
diff changeset
174 Ok((Some(metadata), data))
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
175 } else {
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
176 Err(HgError::corrupted(
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
177 "Missing metadata end delimiter in filelog entry",
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
178 ))
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
179 }
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
180 } else {
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
181 Ok((None, &self.0))
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
182 }
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
183 }
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
184
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
185 /// Returns the file contents at this revision, stripped of any metadata
48569
20d0d896183e rhg: Rename some revlog-related types and methods
Simon Sapin <simon.sapin@octobus.net>
parents: 48260
diff changeset
186 pub fn file_data(&self) -> Result<&[u8], HgError> {
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
187 let (_metadata, data) = self.split()?;
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
188 Ok(data)
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
189 }
48249
027ebad952ac rhg: internally, return a structured representation from hg cat
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 47997
diff changeset
190
027ebad952ac rhg: internally, return a structured representation from hg cat
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 47997
diff changeset
191 /// Consume the entry, and convert it into data, discarding any metadata,
027ebad952ac rhg: internally, return a structured representation from hg cat
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 47997
diff changeset
192 /// if present.
48569
20d0d896183e rhg: Rename some revlog-related types and methods
Simon Sapin <simon.sapin@octobus.net>
parents: 48260
diff changeset
193 pub fn into_file_data(self) -> Result<Vec<u8>, HgError> {
48260
e9faae0f445c rhg: simplify split_metadata
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 48249
diff changeset
194 if let (Some(_metadata), data) = self.split()? {
48249
027ebad952ac rhg: internally, return a structured representation from hg cat
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 47997
diff changeset
195 Ok(data.to_owned())
027ebad952ac rhg: internally, return a structured representation from hg cat
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 47997
diff changeset
196 } else {
027ebad952ac rhg: internally, return a structured representation from hg cat
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 47997
diff changeset
197 Ok(self.0)
027ebad952ac rhg: internally, return a structured representation from hg cat
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 47997
diff changeset
198 }
027ebad952ac rhg: internally, return a structured representation from hg cat
Arseniy Alekseyev <aalekseyev@janestreet.com>
parents: 47997
diff changeset
199 }
47989
4d2a5ca060e3 rust: Add a Filelog struct that wraps Revlog
Simon Sapin <simon.sapin@octobus.net>
parents:
diff changeset
200 }