1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
|
README for Lorry
================
Lorry is a tool to take upstream source code (in various formats,
though preferably in version control) and converts it into a git
repository.
If you want to try this, use `--pull-only` and/or `--mirror-base-url-push`
so that you do not accidentally overwrite important stuff for Baserock.
(If you don't have direct commit access to Baserock on git.baserock.org
then you're not dangerous.)
See the manual page for instructions on using.
The usual use is that a server hosts lorry and checks for updates,
which it then lorries. Lorry should not generally be run from the
developer's machine apart from testing. This is because Lorry has to
keep the git trees to avoid it having to pull everything.
You can find a lot of lorries to crib ideas from at:
git://git.baserock.org/baserock/lorries.git
Implementation
--------------
Lorry relies on git-svn, git-cvsimport, hg-fast-export, perl (for tarballs) and
bzr fast-export for the conversions. You need to have them installed.
Lorry file specification
------------------------
Lorry files are json dicts where the repository names are the keys and the
values are dicts with the data required to mirror it.
So a simple lorry that mirrors a git project looks like
{
"git": {
"type": "git",
"url": "git://github.com/gitster/git.git"
}
}
Multiple repositories can be specified in the same .lorry file, in which case
all of them will be processed by lorry. The following shows two repositories.
{
"git": {
"type": "git",
"url": "git://github.com/gitster/git.git"
},
"curl": {
"type": "git",
"url": "git://github.com/bagder/curl.git"
}
}
Lorry can import other version control systems into git.
### Mercurial
Mercurial is very similar to git, just change the type field to "hg"
{
"sudo": {
"type": "hg",
"url": "http://www.sudo.ws/repos/sudo"
}
}
### Bazaar
Repositories and branches in Bazaar mean different things to Git.
The practical difference for Lorry is that it is not possible to have
a url for a repository, urls map directly to branches.
{
"libpipeline": {
"type": "bzr",
"branches": {
"trunk": "http://bzr.savannah.gnu.org/r/libpipeline/trunk"
}
}
}
For convenience if the project only needs one branch mirrored, the url
is assumed to be the master branch.
{
"libpipeline": {
"type": "bzr",
"url": "http://bzr.savannah.gnu.org/r/libpipeline/trunk"
}
}
### Subversion
To support all the branches and tags a layout needs to be specified as svn is
very flexible with the possible layouts, however the most common is to have the
working branch in a directory called trunk, and the branches and tags in
respectively named subdirectories.
Because this is so common "standard" can be used as the layout
{
"mpc": {
"type": "svn",
"url": "svn://scm.gforge.inria.fr/svn/mpc",
"layout": "standard"
}
}
This is equivalent to
{
"mpc": {
"type": "svn",
"url": "svn://scm.gforge.inria.fr/svn/mpc",
"layout": {
"trunk": "trunk",
"branches": "branches/*",
"tags": "tags/*"
}
}
}
Trunk is the path to the directory where the main branch is located.
Branches and Tags are glob expressions to allow finer control over which paths
are used.
Texlive keeps a lot of resources in their svn repository, we are only concerned
with the source code, so this layout should select the correct subdirectory for
each branch.
{
"texlive": {
"type": "svn",
"url": "svn://tug.org/texlive/",
"layout": {
"trunk": "trunk/Build/source",
"branches": "branches/*/Build/source",
"tags": "tags/*/Build/source"
}
}
}
Brace expansions can be used to specify subsets of paths.
Netpbm for example, keeps all its branches in the root directory
{
"netpbm": {
"type": "svn",
"url": "https://netpbm.svn.sourceforge.net/svnroot/netpbm",
"layout": {
"trunk": "trunk",
"branches": "{advanced,stable,super_stable}",
"tags": "release_number/*"
}
}
}
Note that git-svn can provide better history tracking if the url is as close to
the root of the repository as possible, so it may be more effective if the lorry
was specified similar to this, assuming svnroot is the real root of the repo
{
"netpbm": {
"type": "svn",
"url": "https://netpbm.svn.sourceforge.net/svnroot/",
"layout": {
"trunk": "netpbm/trunk",
"branches": "netpbm/{advanced,stable,super_stable}",
"tags": "netpbm/release_number/*"
}
}
}
### CVS
The url for CVS repositories is the CVSROOT string. The module is required as
cvs repositories usually have multiple modules, the module is usually the same
as the project name.
{
"openssl": {
"type": "cvs",
"url": "anonymous@cvs.openssl.org:/openssl-cvs",
"module": "openssl"
}
}
### Tarball
Lorry can import a tarball fetched from a URL. The contents will be
committed on a branch named after the basename of the tar file (e.g.
bc-1.06.tar.gz will be imported into a branch named 'bc-1.06'.
The import is done by the `lorry.tar-importer` subprocess. It can detect and
handle common compression formats including gzip, bz2, xz and lzma. It will
also detect if there is a 'top directory' that contains the tarball contents
and strip this out of the imported filenames.
Tarball imports once required 'compression' and 'strip' to be specified.
These are obsolete now and are ignored by Lorry.
{
"bc": {
"type": "tarball",
"url": "http://ftp.gnu.org/gnu/bc/bc-1.06.tar.gz"
}
}
NOTE: tarball imports are unlikely to give the same commit SHA1 but the tree
SHA1 inside (which is what is used for artifact cache IDs) should remain
stable.
Tips
----
1. Use upstream's git repository whenever possible
Importing from foreign version control systems is always slower than a
git mirror (with the exception of tarballs because they have less history).
2. GNU Projects often have a git repository
Most GNU projects are old compared to git, so were mainly developed in CVS.
Many official websites only mention the CVS or SVN repositories.
They will tend to have a git repository as well though, especially if they
are hosted on savannah.
Legal stuff
-----------
Copyright (C) 2013 Codethink Limited
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; version 2 of the License.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License along
with this program; if not, write to the Free Software Foundation, Inc.,
51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
|
