1
# Copyright (C) 2007 Canonical Ltd
2
# Copyright (C) 2008-2009 Jelmer Vernooij <jelmer@samba.org>
3
# Copyright (C) 2008 John Carr
5
# This program is free software; you can redistribute it and/or modify
6
# it under the terms of the GNU General Public License as published by
7
# the Free Software Foundation; either version 2 of the License, or
8
# (at your option) any later version.
10
# This program is distributed in the hope that it will be useful,
11
# but WITHOUT ANY WARRANTY; without even the implied warranty of
12
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13
# GNU General Public License for more details.
15
# You should have received a copy of the GNU General Public License
16
# along with this program; if not, write to the Free Software
17
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
19
"""Converters, etc for going between Bazaar and Git ids."""
31
from bzrlib.inventory import (
34
from bzrlib.foreign import (
40
DEFAULT_FILE_MODE = stat.S_IFREG | 0644
43
def escape_file_id(file_id):
44
return file_id.replace('_', '__').replace(' ', '_s')
47
def unescape_file_id(file_id):
50
while i < len(file_id):
52
ret.append(file_id[i])
54
if file_id[i+1] == '_':
56
elif file_id[i+1] == 's':
59
raise AssertionError("unknown escape character %s" % file_id[i+1])
65
def fix_person_identifier(text):
66
if "<" in text and ">" in text:
68
return "%s <%s>" % (text, text)
71
def warn_escaped(commit, num_escaped):
72
trace.warning("Escaped %d XML-invalid characters in %s. Will be unable "
73
"to regenerate the SHA map.", num_escaped, commit)
76
def warn_unusual_mode(commit, path, mode):
77
trace.warning("Unusual file mode %o for %s in %s. Will be unable to "
78
"regenerate the SHA map.", mode, path, commit)
81
def squash_revision(target_repo, rev):
82
"""Remove characters that can't be stored from a revision, if necessary.
84
:param target_repo: Repository in which the revision will be stored
85
:param rev: Revision object, will be modified in-place
87
if not getattr(target_repo._serializer, "squashes_xml_invalid_characters", True):
89
from bzrlib.xml_serializer import escape_invalid_chars
90
rev.message, num_escaped = escape_invalid_chars(rev.message)
92
warn_escaped(rev.foreign_revid, num_escaped)
93
if 'author' in rev.properties:
94
rev.properties['author'], num_escaped = escape_invalid_chars(
95
rev.properties['author'])
97
warn_escaped(rev.foreign_revid, num_escaped)
98
rev.committer, num_escaped = escape_invalid_chars(rev.committer)
100
warn_escaped(rev.foreign_revid, num_escaped)
103
class BzrGitMapping(foreign.VcsMapping):
104
"""Class that maps between Git and Bazaar semantics."""
108
super(BzrGitMapping, self).__init__(foreign_git)
110
def __eq__(self, other):
111
return type(self) == type(other) and self.revid_prefix == other.revid_prefix
114
def revision_id_foreign_to_bzr(cls, git_rev_id):
115
"""Convert a git revision id handle to a Bazaar revision id."""
116
return "%s:%s" % (cls.revid_prefix, git_rev_id)
119
def revision_id_bzr_to_foreign(cls, bzr_rev_id):
120
"""Convert a Bazaar revision id to a git revision id handle."""
121
if not bzr_rev_id.startswith("%s:" % cls.revid_prefix):
122
raise errors.InvalidRevisionId(bzr_rev_id, cls)
123
return bzr_rev_id[len(cls.revid_prefix)+1:], cls()
125
def generate_file_id(self, path):
126
# Git paths are just bytestrings
127
# We must just hope they are valid UTF-8..
130
return escape_file_id(path)
132
def parse_file_id(self, file_id):
133
if file_id == ROOT_ID:
135
return unescape_file_id(file_id)
137
def import_unusual_file_modes(self, rev, unusual_file_modes):
138
if unusual_file_modes:
139
rev.properties['file-modes'] = bencode.bencode(unusual_file_modes)
141
def export_unusual_file_modes(self, rev):
143
return bencode.bdecode(rev.properties['file-modes'])
147
def import_commit(self, commit):
148
"""Convert a git commit to a bzr revision.
150
:return: a `bzrlib.revision.Revision` object.
153
raise AssertionError("Commit object can't be None")
154
rev = ForeignRevision(commit.id, self, self.revision_id_foreign_to_bzr(commit.id))
155
rev.parent_ids = tuple([self.revision_id_foreign_to_bzr(p) for p in commit.parents])
156
rev.message = commit.message.decode("utf-8", "replace")
157
rev.committer = str(commit.committer).decode("utf-8", "replace")
158
if commit.committer != commit.author:
159
rev.properties['author'] = str(commit.author).decode("utf-8", "replace")
161
if commit.commit_time != commit.author_time:
162
rev.properties['author-timestamp'] = str(commit.author_time)
163
if commit.commit_timezone != commit.author_timezone:
164
rev.properties['author-timezone'] = "%d" % (commit.author_timezone, )
165
rev.timestamp = commit.commit_time
166
rev.timezone = commit.commit_timezone
170
class BzrGitMappingv1(BzrGitMapping):
171
revid_prefix = 'git-v1'
175
return self.revid_prefix
178
class BzrGitMappingExperimental(BzrGitMappingv1):
179
revid_prefix = 'git-experimental'
183
class GitMappingRegistry(VcsMappingRegistry):
184
"""Registry with available git mappings."""
186
def revision_id_bzr_to_foreign(self, bzr_revid):
187
if not bzr_revid.startswith("git-"):
188
raise errors.InvalidRevisionId(bzr_revid, None)
189
(mapping_version, git_sha) = bzr_revid.split(":", 1)
190
mapping = self.get(mapping_version)
191
return mapping.revision_id_bzr_to_foreign(bzr_revid)
193
parse_revision_id = revision_id_bzr_to_foreign
196
mapping_registry = GitMappingRegistry()
197
mapping_registry.register_lazy('git-v1', "bzrlib.plugins.git.mapping",
199
mapping_registry.register_lazy('git-experimental', "bzrlib.plugins.git.mapping",
200
"BzrGitMappingExperimental")
203
class ForeignGit(ForeignVcs):
204
"""The Git Stupid Content Tracker"""
207
super(ForeignGit, self).__init__(mapping_registry)
210
def show_foreign_revid(cls, foreign_revid):
211
return { "git commit": foreign_revid }
214
foreign_git = ForeignGit()
215
default_mapping = BzrGitMappingv1()
218
def text_to_blob(texts, entry):
219
from dulwich.objects import Blob
220
text = texts.get_record_stream([(entry.file_id, entry.revision)], 'unordered', True).next().get_bytes_as('fulltext')
226
def symlink_to_blob(entry):
227
from dulwich.objects import Blob
229
blob._text = entry.symlink_target
233
def mode_is_executable(mode):
234
"""Check if mode should be considered executable."""
235
return bool(mode & 0111)
239
"""Determine the Bazaar inventory kind based on Unix file mode."""
240
entry_kind = (mode & 0700000) / 0100000
243
elif entry_kind == 1:
244
file_kind = (mode & 070000) / 010000
250
return 'tree-reference'
252
raise AssertionError(
253
"Unknown file kind %d, perms=%o." % (file_kind, mode,))
255
raise AssertionError(
256
"Unknown kind, perms=%r." % (mode,))
259
def entry_mode(entry):
260
"""Determine the git file mode for an inventory entry."""
261
if entry.kind == 'directory':
263
elif entry.kind == 'symlink':
265
elif entry.kind == 'file':
266
mode = stat.S_IFREG | 0644
274
def directory_to_tree(entry, lookup_ie_sha1):
275
from dulwich.objects import Tree
277
for name in sorted(entry.children.keys()):
278
ie = entry.children[name]
279
tree.add(entry_mode(ie), name.encode("utf-8"), lookup_ie_sha1(ie))
284
def extract_unusual_modes(rev):
286
foreign_revid, mapping = mapping_registry.parse_revision_id(rev.revision_id)
287
except errors.InvalidRevisionId:
290
return mapping.export_unusual_file_modes(rev)
293
def inventory_to_tree_and_blobs(inventory, texts, mapping, cur=None):
294
"""Convert a Bazaar tree to a Git tree.
296
:return: Yields tuples with object sha1, object and path
298
from dulwich.objects import Tree
305
# stack contains the set of trees that we haven't
306
# finished constructing
307
for path, entry in inventory.iter_entries():
308
while stack and not path.startswith(osutils.pathjoin(cur, "")):
309
# We've hit a file that's not a child of the previous path
312
yield sha, tree, cur.encode("utf-8")
313
t = (stat.S_IFDIR, urlutils.basename(cur).encode('UTF-8'), sha)
314
cur, tree = stack.pop()
317
if entry.kind == "directory":
318
stack.append((cur, tree))
322
if entry.kind == "file":
323
blob = text_to_blob(texts, entry)
324
elif entry.kind == "symlink":
325
blob = symlink_to_blob(entry)
327
raise AssertionError("Unknown kind %s" % entry.kind)
329
yield sha, blob, path.encode("utf-8")
330
name = urlutils.basename(path).encode("utf-8")
331
tree.add(entry_mode(entry), name, sha)
333
while len(stack) > 1:
336
yield sha, tree, cur.encode("utf-8")
337
t = (stat.S_IFDIR, urlutils.basename(cur).encode('UTF-8'), sha)
338
cur, tree = stack.pop()
342
yield tree.id, tree, cur.encode("utf-8")
345
def revision_to_commit(rev, tree_sha, parent_lookup):
346
"""Turn a Bazaar revision in to a Git commit
348
:param tree_sha: Tree sha for the commit
349
:param parent_lookup: Function for looking up the GIT sha equiv of a bzr revision
350
:return dulwich.objects.Commit represent the revision:
352
from dulwich.objects import Commit
354
commit.tree = tree_sha
355
for p in rev.parent_ids:
356
git_p = parent_lookup(p)
357
if git_p is not None:
358
assert len(git_p) == 40, "unexpected length for %r" % git_p
359
commit.parents.append(git_p)
360
commit.message = rev.message.encode("utf-8")
361
commit.committer = fix_person_identifier(rev.committer.encode("utf-8"))
362
commit.author = fix_person_identifier(rev.get_apparent_authors()[0].encode("utf-8"))
363
commit.commit_time = long(rev.timestamp)
364
if 'author-timestamp' in rev.properties:
365
commit.author_time = long(rev.properties['author-timestamp'])
367
commit.author_time = commit.commit_time
368
commit.commit_timezone = rev.timezone
369
if 'author-timezone' in rev.properties:
370
commit.author_timezone = int(rev.properties['author-timezone'])
372
commit.author_timezone = commit.commit_timezone
added
removed
mapping.py
