annotate hgext/git/index.py @ 52624:cdbfe5e7592e

git: handle octopus merges Octopus merges in git are merge commits with more than 2 parents. To make them fit into mercurial core's assumption about commits having 0-2 parents, the git indexing code creates "sythetic" commits to represent the octopus commit as a sequence of regular 2-parent commits. The synthetic commit hashes are just an incrementing commit number (which is the same as the generated rev number). The last commit in the sequence of commits uses the actual git commit hash. As a result, `hg checkout -r <commit>` produces the same working directory as `git checkout <commit>` for all git commit hashes. The synthetic commit hashes are stored in the changelog table as any other commit - with the two parents - but they also contain the commit hash of the octopus merge commit. For example, given the git DAG (manually pruned `git log --graph`): *-. commit 23480d86e2689703b33f693907c40fbe6e1620e4 Merge branches... |\ \ | | | | | * commit 2eda9984b06c75448598ec6c0a9028e49dacf616 C | | | | * | commit 5e634a12f12fedaf7b8ef0f0fcdbb07222871953 B | |/ | | * | commit 8883a1296c5ae323a1b18d1f6410398ce43ebd3a D |/ | * commit 95f241588fded9554cae91be0fefd576f61ebfc6 A Where M is the octopus merge commit with 3 parents, the corresponding mercurial DAG is: $ hg log -G -T '{node} {desc}' @ 23480d86e2689703b33f693907c40fbe6e1620e4 Merge branches 'abc' and 'def' |\ | o 0000000000000000000000000000000000000004 Merge branches 'abc' and 'def' | |\ | | o 8883a1296c5ae323a1b18d1f6410398ce43ebd3a D | | | o---+ 2eda9984b06c75448598ec6c0a9028e49dacf616 C / / o / 5e634a12f12fedaf7b8ef0f0fcdbb07222871953 B |/ o 95f241588fded9554cae91be0fefd576f61ebfc6 A
author Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
date Sun, 10 Mar 2024 14:30:32 -0400
parents 4e2ea270ba6a
children 27a0bfe770eb
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
51859
f4733654f144 typing: add `from __future__ import annotations` to most files
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
1 from __future__ import annotations
f4733654f144 typing: add `from __future__ import annotations` to most files
Matt Harbison <matt_harbison@yahoo.com>
parents: 48875
diff changeset
2
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
3 import collections
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
4 import os
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
5 import sqlite3
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
6
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
7 from mercurial.i18n import _
47012
d55b71393907 node: replace nullid and friends with nodeconstants class
Joerg Sonnenberger <joerg@bec.de>
parents: 46113
diff changeset
8 from mercurial.node import sha1nodeconstants
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
9
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
10 from mercurial import (
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
11 encoding,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
12 error,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
13 pycompat,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
14 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
15
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
16 from . import gitutil
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
17
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
18
44484
ec54b3d2af0b git: don't fail import when pygit2 is not install
Martin von Zweigbergk <martinvonz@google.com>
parents: 44477
diff changeset
19 pygit2 = gitutil.get_pygit2()
ec54b3d2af0b git: don't fail import when pygit2 is not install
Martin von Zweigbergk <martinvonz@google.com>
parents: 44477
diff changeset
20
52624
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
21 _CURRENT_SCHEMA_VERSION = 3
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
22 _SCHEMA = (
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
23 """
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
24 CREATE TABLE refs (
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
25 -- node and name are unique together. There may be more than one name for
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
26 -- a given node, and there may be no name at all for a given node (in the
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
27 -- case of an anonymous hg head).
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
28 node TEXT NOT NULL,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
29 name TEXT
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
30 );
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
31
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
32 -- The "possible heads" of the repository, which we use to figure out
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
33 -- if we need to re-walk the changelog.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
34 CREATE TABLE possible_heads (
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
35 node TEXT NOT NULL
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
36 );
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
37
52622
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
38 CREATE UNIQUE INDEX possible_heads_idx ON possible_heads(node);
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
39
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
40 -- The topological heads of the changelog, which hg depends on.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
41 CREATE TABLE heads (
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
42 node TEXT NOT NULL
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
43 );
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
44
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
45 -- A total ordering of the changelog
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
46 CREATE TABLE changelog (
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
47 rev INTEGER NOT NULL PRIMARY KEY,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
48 node TEXT NOT NULL,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
49 p1 TEXT,
52624
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
50 p2 TEXT,
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
51 synthetic TEXT
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
52 );
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
53
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
54 CREATE UNIQUE INDEX changelog_node_idx ON changelog(node);
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
55 CREATE UNIQUE INDEX changelog_node_rev_idx ON changelog(rev, node);
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
56
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
57 -- Changed files for each commit, which lets us dynamically build
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
58 -- filelogs.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
59 CREATE TABLE changedfiles (
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
60 node TEXT NOT NULL,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
61 filename TEXT NOT NULL,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
62 -- 40 zeroes for deletions
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
63 filenode TEXT NOT NULL,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
64 -- to handle filelog parentage:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
65 p1node TEXT,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
66 p1filenode TEXT,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
67 p2node TEXT,
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
68 p2filenode TEXT
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
69 );
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
70
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
71 CREATE INDEX changedfiles_nodes_idx
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
72 ON changedfiles(node);
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
73
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
74 PRAGMA user_version=%d
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
75 """
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
76 % _CURRENT_SCHEMA_VERSION
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
77 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
78
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
79
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
80 def _createdb(path):
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
81 # print('open db', path)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
82 # import traceback
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
83 # traceback.print_stack()
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
84 db = sqlite3.connect(encoding.strfromlocal(path))
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
85 db.text_factory = bytes
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
86
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
87 res = db.execute('PRAGMA user_version').fetchone()[0]
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
88
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
89 # New database.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
90 if res == 0:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
91 for statement in _SCHEMA.split(';'):
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
92 db.execute(statement.strip())
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
93
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
94 db.commit()
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
95
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
96 elif res == _CURRENT_SCHEMA_VERSION:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
97 pass
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
98
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
99 else:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
100 raise error.Abort(_(b'sqlite database has unrecognized version'))
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
101
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
102 db.execute('PRAGMA journal_mode=WAL')
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
103
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
104 return db
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
105
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
106
44484
ec54b3d2af0b git: don't fail import when pygit2 is not install
Martin von Zweigbergk <martinvonz@google.com>
parents: 44477
diff changeset
107 _OUR_ORDER = ()
ec54b3d2af0b git: don't fail import when pygit2 is not install
Martin von Zweigbergk <martinvonz@google.com>
parents: 44477
diff changeset
108 if pygit2:
ec54b3d2af0b git: don't fail import when pygit2 is not install
Martin von Zweigbergk <martinvonz@google.com>
parents: 44477
diff changeset
109 _OUR_ORDER = (
ec54b3d2af0b git: don't fail import when pygit2 is not install
Martin von Zweigbergk <martinvonz@google.com>
parents: 44477
diff changeset
110 pygit2.GIT_SORT_TOPOLOGICAL
ec54b3d2af0b git: don't fail import when pygit2 is not install
Martin von Zweigbergk <martinvonz@google.com>
parents: 44477
diff changeset
111 | pygit2.GIT_SORT_TIME
ec54b3d2af0b git: don't fail import when pygit2 is not install
Martin von Zweigbergk <martinvonz@google.com>
parents: 44477
diff changeset
112 | pygit2.GIT_SORT_REVERSE
ec54b3d2af0b git: don't fail import when pygit2 is not install
Martin von Zweigbergk <martinvonz@google.com>
parents: 44477
diff changeset
113 )
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
114
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
115 _DIFF_FLAGS = 1 << 21 # GIT_DIFF_FORCE_BINARY, which isn't exposed by pygit2
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
116
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
117
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
118 def _find_nearest_ancestor_introducing_node(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
119 db, gitrepo, file_path, walk_start, filenode
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
120 ):
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
121 """Find the nearest ancestor that introduces a file node.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
122
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
123 Args:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
124 db: a handle to our sqlite database.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
125 gitrepo: A pygit2.Repository instance.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
126 file_path: the path of a file in the repo
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
127 walk_start: a pygit2.Oid that is a commit where we should start walking
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
128 for our nearest ancestor.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
129
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
130 Returns:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
131 A hexlified SHA that is the commit ID of the next-nearest parent.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
132 """
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
133 assert isinstance(file_path, str), 'file_path must be str, got %r' % type(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
134 file_path
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
135 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
136 assert isinstance(filenode, str), 'filenode must be str, got %r' % type(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
137 filenode
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
138 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
139 parent_options = {
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
140 row[0].decode('ascii')
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
141 for row in db.execute(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
142 'SELECT node FROM changedfiles '
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
143 'WHERE filename = ? AND filenode = ?',
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
144 (file_path, filenode),
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
145 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
146 }
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
147 inner_walker = gitrepo.walk(walk_start, _OUR_ORDER)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
148 for w in inner_walker:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
149 if w.id.hex in parent_options:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
150 return w.id.hex
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
151 raise error.ProgrammingError(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
152 'Unable to find introducing commit for %s node %s from %s',
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
153 (file_path, filenode, walk_start),
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
154 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
155
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
156
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
157 def fill_in_filelog(gitrepo, db, startcommit, path, startfilenode):
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
158 """Given a starting commit and path, fill in a filelog's parent pointers.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
159
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
160 Args:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
161 gitrepo: a pygit2.Repository
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
162 db: a handle to our sqlite database
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
163 startcommit: a hexlified node id for the commit to start at
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
164 path: the path of the file whose parent pointers we should fill in.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
165 filenode: the hexlified node id of the file at startcommit
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
166
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
167 TODO: make filenode optional
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
168 """
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
169 assert isinstance(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
170 startcommit, str
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
171 ), 'startcommit must be str, got %r' % type(startcommit)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
172 assert isinstance(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
173 startfilenode, str
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
174 ), 'startfilenode must be str, got %r' % type(startfilenode)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
175 visit = collections.deque([(startcommit, startfilenode)])
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
176 while visit:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
177 cnode, filenode = visit.popleft()
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
178 commit = gitrepo[cnode]
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
179 parents = []
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
180 for parent in commit.parents:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
181 t = parent.tree
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
182 for comp in path.split('/'):
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
183 try:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
184 t = gitrepo[t[comp].id]
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
185 except KeyError:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
186 break
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
187 else:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
188 introducer = _find_nearest_ancestor_introducing_node(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
189 db, gitrepo, path, parent.id, t.id.hex
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
190 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
191 parents.append((introducer, t.id.hex))
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
192 p1node = p1fnode = p2node = p2fnode = gitutil.nullgit
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
193 for par, parfnode in parents:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
194 found = int(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
195 db.execute(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
196 'SELECT COUNT(*) FROM changedfiles WHERE '
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
197 'node = ? AND filename = ? AND filenode = ? AND '
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
198 'p1node NOT NULL',
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
199 (par, path, parfnode),
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
200 ).fetchone()[0]
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
201 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
202 if found == 0:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
203 assert par is not None
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
204 visit.append((par, parfnode))
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
205 if parents:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
206 p1node, p1fnode = parents[0]
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
207 if len(parents) == 2:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
208 p2node, p2fnode = parents[1]
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
209 if len(parents) > 2:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
210 raise error.ProgrammingError(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
211 b"git support can't handle octopus merges"
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
212 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
213 db.execute(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
214 'UPDATE changedfiles SET '
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
215 'p1node = ?, p1filenode = ?, p2node = ?, p2filenode = ? '
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
216 'WHERE node = ? AND filename = ? AND filenode = ?',
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
217 (p1node, p1fnode, p2node, p2fnode, commit.id.hex, path, filenode),
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
218 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
219 db.commit()
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
220
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
221
52623
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
222 def _index_repo_commit(gitrepo, db, commit):
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
223 files = {}
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
224 # I *think* we only need to check p1 for changed files
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
225 # (and therefore linkrevs), because any node that would
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
226 # actually have this commit as a linkrev would be
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
227 # completely new in this rev.
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
228 p1 = commit.parents[0].id.hex if commit.parents else None
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
229 if p1 is not None:
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
230 patchgen = gitrepo.diff(p1, commit.id.hex, flags=_DIFF_FLAGS)
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
231 else:
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
232 patchgen = commit.tree.diff_to_tree(swap=True, flags=_DIFF_FLAGS)
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
233 new_files = (p.delta.new_file for p in patchgen)
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
234 files = {
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
235 nf.path: nf.id.hex
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
236 for nf in new_files
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
237 if nf.id.raw != sha1nodeconstants.nullid
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
238 }
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
239 for p, n in files.items():
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
240 # We intentionally set NULLs for any file parentage
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
241 # information so it'll get demand-computed later. We
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
242 # used to do it right here, and it was _very_ slow.
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
243 db.execute(
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
244 'INSERT INTO changedfiles ('
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
245 'node, filename, filenode, p1node, p1filenode, p2node, '
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
246 'p2filenode) VALUES(?, ?, ?, ?, ?, ?, ?)',
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
247 (commit.id.hex, p, n, None, None, None, None),
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
248 )
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
249
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
250
44951
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
251 def _index_repo(
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
252 gitrepo,
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
253 db,
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
254 logfn=lambda x: None,
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
255 progress_factory=lambda *args, **kwargs: None,
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
256 ):
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
257 # Identify all references so we can tell the walker to visit all of them.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
258 all_refs = gitrepo.listall_references()
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
259 possible_heads = set()
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
260 prog = progress_factory(b'refs')
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
261 for pos, ref in enumerate(all_refs):
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
262 if prog is not None:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
263 prog.update(pos)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
264 if not (
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
265 ref.startswith('refs/heads/') # local branch
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
266 or ref.startswith('refs/tags/') # tag
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
267 or ref.startswith('refs/remotes/') # remote branch
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
268 or ref.startswith('refs/hg/') # from this extension
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
269 ):
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
270 continue
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
271 try:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
272 start = gitrepo.lookup_reference(ref).peel(pygit2.GIT_OBJ_COMMIT)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
273 except ValueError:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
274 # No commit to be found, so we don't care for hg's purposes.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
275 continue
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
276 possible_heads.add(start.id)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
277 # Optimization: if the list of heads hasn't changed, don't
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
278 # reindex, the changelog. This doesn't matter on small
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
279 # repositories, but on even moderately deep histories (eg cpython)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
280 # this is a very important performance win.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
281 #
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
282 # TODO: we should figure out how to incrementally index history
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
283 # (preferably by detecting rewinds!) so that we don't have to do a
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
284 # full changelog walk every time a new commit is created.
44946
fb2936c5f6dc git: decode node IDs back into Python strings (issue6349)
Hollis Blanchard <hollis_blanchard@mentor.com>
parents: 44484
diff changeset
285 cache_heads = {
fb2936c5f6dc git: decode node IDs back into Python strings (issue6349)
Hollis Blanchard <hollis_blanchard@mentor.com>
parents: 44484
diff changeset
286 pycompat.sysstr(x[0])
fb2936c5f6dc git: decode node IDs back into Python strings (issue6349)
Hollis Blanchard <hollis_blanchard@mentor.com>
parents: 44484
diff changeset
287 for x in db.execute('SELECT node FROM possible_heads')
fb2936c5f6dc git: decode node IDs back into Python strings (issue6349)
Hollis Blanchard <hollis_blanchard@mentor.com>
parents: 44484
diff changeset
288 }
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
289 walker = None
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
290 cur_cache_heads = {h.hex for h in possible_heads}
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
291 if cur_cache_heads == cache_heads:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
292 return
44951
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
293 logfn(b'heads mismatch, rebuilding dagcache\n')
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
294 for start in possible_heads:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
295 if walker is None:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
296 walker = gitrepo.walk(start, _OUR_ORDER)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
297 else:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
298 walker.push(start)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
299
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
300 # Empty out the existing changelog. Even for large-ish histories
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
301 # we can do the top-level "walk all the commits" dance very
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
302 # quickly as long as we don't need to figure out the changed files
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
303 # list.
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
304 db.execute('DELETE FROM changelog')
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
305 if prog is not None:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
306 prog.complete()
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
307 prog = progress_factory(b'commits')
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
308 # This walker is sure to visit all the revisions in history, but
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
309 # only once.
52624
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
310 pos = -1
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
311 for commit in walker:
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
312 if prog is not None:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
313 prog.update(pos)
47051
de26b9a7ec29 git: consistently use str for parents when rebuilding the index database
Matt Harbison <matt_harbison@yahoo.com>
parents: 46113
diff changeset
314 p1 = p2 = gitutil.nullgit
52624
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
315 if len(commit.parents) <= 2:
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
316 if commit.parents:
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
317 p1 = commit.parents[0].id.hex
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
318 if len(commit.parents) == 2:
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
319 p2 = commit.parents[1].id.hex
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
320 pos += 1
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
321 db.execute(
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
322 'INSERT INTO changelog (rev, node, p1, p2, synthetic) VALUES(?, ?, ?, ?, NULL)',
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
323 (pos, commit.id.hex, p1, p2),
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
324 )
52624
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
325 else:
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
326 parents = list(commit.parents)
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
327
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
328 p1 = parents.pop(0).id.hex
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
329 while parents:
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
330 pos += 1
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
331
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
332 if len(parents) == 1:
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
333 this = commit.id.hex
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
334 synth = None
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
335 else:
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
336 this = "%040x" % pos
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
337 synth = commit.id.hex
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
338
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
339 p2 = parents.pop(0).id.hex
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
340
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
341 db.execute(
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
342 'INSERT INTO changelog (rev, node, p1, p2, synthetic) VALUES(?, ?, ?, ?, ?)',
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
343 (pos, this, p1, p2, synth),
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
344 )
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
345
cdbfe5e7592e git: handle octopus merges
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52623
diff changeset
346 p1 = this
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
347
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
348 num_changedfiles = db.execute(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
349 "SELECT COUNT(*) from changedfiles WHERE node = ?",
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
350 (commit.id.hex,),
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
351 ).fetchone()[0]
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
352 if not num_changedfiles:
52623
4e2ea270ba6a git: move file indexing into a helper
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 52622
diff changeset
353 _index_repo_commit(gitrepo, db, commit)
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
354 db.execute('DELETE FROM heads')
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
355 db.execute('DELETE FROM possible_heads')
52622
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
356 db.executemany(
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
357 'INSERT INTO possible_heads (node) VALUES(?)',
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
358 [(hid.hex,) for hid in possible_heads],
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
359 )
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
360 db.execute(
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
361 '''
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
362 INSERT INTO heads (node)
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
363 SELECT node FROM possible_heads WHERE
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
364 node NOT IN (
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
365 SELECT DISTINCT possible_heads.node FROM changelog, possible_heads WHERE
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
366 changelog.p1 = possible_heads.node OR
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
367 changelog.p2 = possible_heads.node
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
368 )
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
369 '''
aa5844ade247 git: speed up possible head processing during indexing by ~100x
Josef 'Jeff' Sipek <jeffpc@josefsipek.net>
parents: 51859
diff changeset
370 )
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
371
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
372 db.commit()
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
373 if prog is not None:
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
374 prog.complete()
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
375
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
376
44951
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
377 def get_index(
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
378 gitrepo, logfn=lambda x: None, progress_factory=lambda *args, **kwargs: None
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
379 ):
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
380 cachepath = os.path.join(
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
381 pycompat.fsencode(gitrepo.path), b'..', b'.hg', b'cache'
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
382 )
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
383 if not os.path.exists(cachepath):
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
384 os.makedirs(cachepath)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
385 dbpath = os.path.join(cachepath, b'git-commits.sqlite')
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
386 db = _createdb(dbpath)
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
387 # TODO check against gitrepo heads before doing a full index
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
388 # TODO thread a ui.progress call into this layer
44951
83e41b73d115 git: add debug logging when there's a mismatch in the cached heads list
Augie Fackler <augie@google.com>
parents: 44946
diff changeset
389 _index_repo(gitrepo, db, logfn, progress_factory)
44477
ad718271a9eb git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff changeset
390 return db