Mercurial > public > mercurial-scm > hg
annotate hgext/git/index.py @ 44477:ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
This is based in part of work I did years ago in hgit, but it's mostly
new code since I'm using pygit2 instead of dulwich and the hg storage
interfaces have improved. Some cleanup of old hgit code by Pulkit,
which I greatly appreciate.
test-git-interop.t does not cover a whole lot of cases, but it
passes. It includes status, diff, making a new commit, and `hg annotate`
working on the git repository.
This is _not_ (yet) production quality code: this is an
experiment. Known technical debt lurking in this implementation:
* Writing bookmarks just totally ignores transactions.
* The way progress is threaded down into the gitstore is awful.
* Ideally we'd find a way to incrementally reindex DAGs. I'm not sure
how to do that efficiently, so we might need a "known only fast-forwards"
mode on the DAG indexer for use on `hg commit` and friends.
* We don't even _try_ to do anything reasonable for `hg pull` or `hg push`.
* Mercurial need an interface for the changelog type.
Tests currently require git 2.24 as far as I'm aware: `git status` has
some changed output that I didn't try and handle in a compatible way.
This patch has produced some interesting cleanups, most recently on
the manifest type. I expect continuing down this road will produce
other meritorious cleanups throughout our code.
Differential Revision: https://phab.mercurial-scm.org/D6734
author | Augie Fackler <augie@google.com> |
---|---|
date | Tue, 11 Feb 2020 00:44:59 -0500 |
parents | |
children | ec54b3d2af0b |
rev | line source |
---|---|
44477
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
1 from __future__ import absolute_import |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
2 |
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 import pygit2 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
8 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
9 from mercurial.i18n import _ |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
10 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
11 from mercurial import ( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
12 encoding, |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
13 error, |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
14 node as nodemod, |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
15 pycompat, |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
16 ) |
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 from . import gitutil |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
19 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
20 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
21 _CURRENT_SCHEMA_VERSION = 1 |
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 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
38 -- 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
|
39 CREATE TABLE heads ( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
40 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
|
41 ); |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
42 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
43 -- 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
|
44 CREATE TABLE changelog ( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
45 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
|
46 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
|
47 p1 TEXT, |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
48 p2 TEXT |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
49 ); |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
50 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
51 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
|
52 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
|
53 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
54 -- 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
|
55 -- filelogs. |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
56 CREATE TABLE changedfiles ( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
57 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
|
58 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
|
59 -- 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
|
60 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
|
61 -- 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
|
62 p1node TEXT, |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
63 p1filenode TEXT, |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
64 p2node TEXT, |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
65 p2filenode TEXT |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
66 ); |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
67 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
68 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
|
69 ON changedfiles(node); |
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 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
|
72 """ |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
73 % _CURRENT_SCHEMA_VERSION |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
74 ) |
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 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
77 def _createdb(path): |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
78 # 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
|
79 # import traceback |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
80 # traceback.print_stack() |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
81 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
|
82 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
|
83 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
84 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
|
85 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
86 # New database. |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
87 if res == 0: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
88 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
|
89 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
|
90 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
91 db.commit() |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
92 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
93 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
|
94 pass |
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 else: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
97 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
|
98 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
99 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
|
100 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
101 return db |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
102 |
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 _OUR_ORDER = ( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
105 pygit2.GIT_SORT_TOPOLOGICAL | pygit2.GIT_SORT_TIME | pygit2.GIT_SORT_REVERSE |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
106 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
107 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
108 _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
|
109 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
110 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
111 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
|
112 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
|
113 ): |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
114 """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
|
115 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
116 Args: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
117 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
|
118 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
|
119 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
|
120 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
|
121 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
|
122 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
123 Returns: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
124 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
|
125 """ |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
126 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
|
127 file_path |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
128 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
129 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
|
130 filenode |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
131 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
132 parent_options = { |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
133 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
|
134 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
|
135 '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
|
136 '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
|
137 (file_path, 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 } |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
140 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
|
141 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
|
142 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
|
143 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
|
144 raise error.ProgrammingError( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
145 '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
|
146 (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
|
147 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
148 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
149 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
150 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
|
151 """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
|
152 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
153 Args: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
154 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
|
155 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
|
156 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
|
157 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
|
158 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
|
159 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
160 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
|
161 """ |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
162 assert isinstance( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
163 startcommit, str |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
164 ), '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
|
165 assert isinstance( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
166 startfilenode, str |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
167 ), '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
|
168 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
|
169 while visit: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
170 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
|
171 commit = gitrepo[cnode] |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
172 parents = [] |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
173 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
|
174 t = parent.tree |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
175 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
|
176 try: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
177 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
|
178 except KeyError: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
179 break |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
180 else: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
181 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
|
182 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
|
183 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
184 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
|
185 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
|
186 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
|
187 found = int( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
188 db.execute( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
189 '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
|
190 '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
|
191 'p1node NOT NULL', |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
192 (par, path, parfnode), |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
193 ).fetchone()[0] |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
194 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
195 if found == 0: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
196 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
|
197 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
|
198 if parents: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
199 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
|
200 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
|
201 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
|
202 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
|
203 raise error.ProgrammingError( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
204 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
|
205 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
206 db.execute( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
207 'UPDATE changedfiles SET ' |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
208 '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
|
209 '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
|
210 (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
|
211 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
212 db.commit() |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
213 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
214 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
215 def _index_repo(gitrepo, db, progress_factory=lambda *args, **kwargs: None): |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
216 # 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
|
217 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
|
218 possible_heads = set() |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
219 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
|
220 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
|
221 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
|
222 prog.update(pos) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
223 if not ( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
224 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
|
225 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
|
226 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
|
227 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
|
228 ): |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
229 continue |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
230 try: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
231 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
|
232 except ValueError: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
233 # 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
|
234 continue |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
235 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
|
236 # 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
|
237 # 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
|
238 # 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
|
239 # 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
|
240 # |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
241 # 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
|
242 # (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
|
243 # full changelog walk every time a new commit is created. |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
244 cache_heads = {x[0] for x in db.execute('SELECT node FROM possible_heads')} |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
245 walker = None |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
246 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
|
247 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
|
248 return |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
249 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
|
250 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
|
251 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
|
252 else: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
253 walker.push(start) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
254 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
255 # 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
|
256 # 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
|
257 # 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
|
258 # list. |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
259 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
|
260 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
|
261 prog.complete() |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
262 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
|
263 # 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
|
264 # only once. |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
265 for pos, commit in enumerate(walker): |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
266 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
|
267 prog.update(pos) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
268 p1 = p2 = nodemod.nullhex |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
269 if len(commit.parents) > 2: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
270 raise error.ProgrammingError( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
271 ( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
272 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
|
273 b"found a commit with %d parents :(" |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
274 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
275 % len(commit.parents) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
276 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
277 if commit.parents: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
278 p1 = commit.parents[0].id.hex |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
279 if len(commit.parents) == 2: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
280 p2 = commit.parents[1].id.hex |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
281 db.execute( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
282 'INSERT INTO changelog (rev, node, p1, p2) VALUES(?, ?, ?, ?)', |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
283 (pos, commit.id.hex, p1, p2), |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
284 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
285 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
286 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
|
287 "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
|
288 (commit.id.hex,), |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
289 ).fetchone()[0] |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
290 if not num_changedfiles: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
291 files = {} |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
292 # I *think* we only need to check p1 for changed files |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
293 # (and therefore linkrevs), because any node that would |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
294 # actually have this commit as a linkrev would be |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
295 # completely new in this rev. |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
296 p1 = commit.parents[0].id.hex if commit.parents else None |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
297 if p1 is not None: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
298 patchgen = gitrepo.diff(p1, commit.id.hex, flags=_DIFF_FLAGS) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
299 else: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
300 patchgen = commit.tree.diff_to_tree( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
301 swap=True, flags=_DIFF_FLAGS |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
302 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
303 new_files = (p.delta.new_file for p in patchgen) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
304 files = { |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
305 nf.path: nf.id.hex |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
306 for nf in new_files |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
307 if nf.id.raw != nodemod.nullid |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
308 } |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
309 for p, n in files.items(): |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
310 # We intentionally set NULLs for any file parentage |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
311 # information so it'll get demand-computed later. We |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
312 # used to do it right here, and it was _very_ slow. |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
313 db.execute( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
314 'INSERT INTO changedfiles (' |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
315 'node, filename, filenode, p1node, p1filenode, p2node, ' |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
316 'p2filenode) VALUES(?, ?, ?, ?, ?, ?, ?)', |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
317 (commit.id.hex, p, n, None, None, None, None), |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
318 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
319 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
|
320 db.execute('DELETE FROM possible_heads') |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
321 for hid in possible_heads: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
322 h = hid.hex |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
323 db.execute('INSERT INTO possible_heads (node) VALUES(?)', (h,)) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
324 haschild = db.execute( |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
325 'SELECT COUNT(*) FROM changelog WHERE p1 = ? OR p2 = ?', (h, h) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
326 ).fetchone()[0] |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
327 if not haschild: |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
328 db.execute('INSERT INTO heads (node) VALUES(?)', (h,)) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
329 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
330 db.commit() |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
331 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
|
332 prog.complete() |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
333 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
334 |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
335 def get_index(gitrepo, progress_factory=lambda *args, **kwargs: None): |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
336 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
|
337 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
|
338 ) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
339 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
|
340 os.makedirs(cachepath) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
341 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
|
342 db = _createdb(dbpath) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
343 # 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
|
344 # TODO thread a ui.progress call into this layer |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
345 _index_repo(gitrepo, db, progress_factory) |
ad718271a9eb
git: skeleton of a new extension to _directly_ operate on git repos
Augie Fackler <augie@google.com>
parents:
diff
changeset
|
346 return db |