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."""
30
from bzrlib import bencode
32
from bzrlib.util import bencode
33
from bzrlib.inventory import (
36
from bzrlib.foreign import (
41
from bzrlib.revision import (
44
from bzrlib.plugins.git.hg import (
49
DEFAULT_FILE_MODE = stat.S_IFREG | 0644
52
def escape_file_id(file_id):
53
return file_id.replace('_', '__').replace(' ', '_s')
56
def unescape_file_id(file_id):
59
while i < len(file_id):
61
ret.append(file_id[i])
63
if file_id[i+1] == '_':
65
elif file_id[i+1] == 's':
68
raise AssertionError("unknown escape character %s" %
75
def fix_person_identifier(text):
76
if "<" in text and ">" in text:
78
return "%s <%s>" % (text, text)
81
def warn_escaped(commit, num_escaped):
82
trace.warning("Escaped %d XML-invalid characters in %s. Will be unable "
83
"to regenerate the SHA map.", num_escaped, commit)
86
def warn_unusual_mode(commit, path, mode):
87
trace.mutter("Unusual file mode %o for %s in %s. Storing as revision "
88
"property. ", mode, path, commit)
91
def squash_revision(target_repo, rev):
92
"""Remove characters that can't be stored from a revision, if necessary.
94
:param target_repo: Repository in which the revision will be stored
95
:param rev: Revision object, will be modified in-place
97
if not getattr(target_repo._serializer, "squashes_xml_invalid_characters", True):
99
from bzrlib.xml_serializer import escape_invalid_chars
100
rev.message, num_escaped = escape_invalid_chars(rev.message)
102
warn_escaped(rev.foreign_revid, num_escaped)
103
if 'author' in rev.properties:
104
rev.properties['author'], num_escaped = escape_invalid_chars(
105
rev.properties['author'])
107
warn_escaped(rev.foreign_revid, num_escaped)
108
rev.committer, num_escaped = escape_invalid_chars(rev.committer)
110
warn_escaped(rev.foreign_revid, num_escaped)
113
class BzrGitMapping(foreign.VcsMapping):
114
"""Class that maps between Git and Bazaar semantics."""
118
super(BzrGitMapping, self).__init__(foreign_git)
120
def __eq__(self, other):
121
return (type(self) == type(other) and
122
self.revid_prefix == other.revid_prefix)
125
def revision_id_foreign_to_bzr(cls, git_rev_id):
126
"""Convert a git revision id handle to a Bazaar revision id."""
127
from dulwich.protocol import ZERO_SHA
128
if git_rev_id == ZERO_SHA:
130
return "%s:%s" % (cls.revid_prefix, git_rev_id)
133
def revision_id_bzr_to_foreign(cls, bzr_rev_id):
134
"""Convert a Bazaar revision id to a git revision id handle."""
135
if not bzr_rev_id.startswith("%s:" % cls.revid_prefix):
136
raise errors.InvalidRevisionId(bzr_rev_id, cls)
137
return bzr_rev_id[len(cls.revid_prefix)+1:], cls()
139
def generate_file_id(self, path):
140
# Git paths are just bytestrings
141
# We must just hope they are valid UTF-8..
144
return escape_file_id(path)
146
def parse_file_id(self, file_id):
147
if file_id == ROOT_ID:
149
return unescape_file_id(file_id)
151
def import_unusual_file_modes(self, rev, unusual_file_modes):
152
if unusual_file_modes:
153
ret = [(path, unusual_file_modes[path])
154
for path in sorted(unusual_file_modes.keys())]
155
rev.properties['file-modes'] = bencode.bencode(ret)
157
def export_unusual_file_modes(self, rev):
159
return dict([(self.generate_file_id(path), mode) for (path, mode) in bencode.bdecode(rev.properties['file-modes'].encode("utf-8"))])
163
def _generate_git_svn_metadata(self, rev, encoding):
165
return "\ngit-svn-id: %s\n" % rev.properties["git-svn-id"].encode(
170
def _generate_hg_message_tail(self, rev):
174
for name in rev.properties:
175
if name == 'hg:extra:branch':
176
branch = rev.properties['hg:extra:branch']
177
elif name.startswith('hg:extra'):
178
extra[name[len('hg:extra:'):]] = base64.b64decode(
179
rev.properties[name])
180
elif name == 'hg:renames':
181
renames = bencode.bdecode(base64.b64decode(
182
rev.properties['hg:renames']))
183
# TODO: Export other properties as 'bzr:' extras?
184
ret = format_hg_metadata(renames, branch, extra)
185
assert isinstance(ret, str)
188
def _extract_git_svn_metadata(self, rev, message):
189
lines = message.split("\n")
190
if not (lines[-1] == "" and lines[-2].startswith("git-svn-id:")):
192
git_svn_id = lines[-2].split(": ", 1)[1]
193
rev.properties['git-svn-id'] = git_svn_id
194
(url, rev, uuid) = parse_git_svn_id(git_svn_id)
195
# FIXME: Convert this to converted-from property somehow..
196
ret = "\n".join(lines[:-2])
197
assert isinstance(ret, str)
200
def _extract_hg_metadata(self, rev, message):
201
(message, renames, branch, extra) = extract_hg_metadata(message)
202
if branch is not None:
203
rev.properties['hg:extra:branch'] = branch
204
for name, value in extra.iteritems():
205
rev.properties['hg:extra:' + name] = base64.b64encode(value)
207
rev.properties['hg:renames'] = base64.b64encode(bencode.bencode(
208
[(new, old) for (old, new) in renames.iteritems()]))
211
def _decode_commit_message(self, rev, message, encoding):
212
return message.decode(encoding)
214
def _encode_commit_message(self, rev, message, encoding):
215
return message.encode(encoding)
217
def export_commit(self, rev, tree_sha, parent_lookup):
218
"""Turn a Bazaar revision in to a Git commit
220
:param tree_sha: Tree sha for the commit
221
:param parent_lookup: Function for looking up the GIT sha equiv of a
223
:return dulwich.objects.Commit represent the revision:
225
from dulwich.objects import Commit
227
commit.tree = tree_sha
228
for p in rev.parent_ids:
230
git_p = parent_lookup(p)
233
if git_p is not None:
234
assert len(git_p) == 40, "unexpected length for %r" % git_p
235
commit.parents.append(git_p)
237
encoding = rev.properties['git-explicit-encoding']
239
encoding = rev.properties.get('git-implicit-encoding', 'utf-8')
240
commit.encoding = rev.properties.get('git-explicit-encoding')
241
commit.committer = fix_person_identifier(rev.committer.encode(
243
commit.author = fix_person_identifier(
244
rev.get_apparent_authors()[0].encode(encoding))
245
commit.commit_time = long(rev.timestamp)
246
if 'author-timestamp' in rev.properties:
247
commit.author_time = long(rev.properties['author-timestamp'])
249
commit.author_time = commit.commit_time
250
commit._commit_timezone_neg_utc = "commit-timezone-neg-utc" in rev.properties
251
commit.commit_timezone = rev.timezone
252
commit._author_timezone_neg_utc = "author-timezone-neg-utc" in rev.properties
253
if 'author-timezone' in rev.properties:
254
commit.author_timezone = int(rev.properties['author-timezone'])
256
commit.author_timezone = commit.commit_timezone
257
commit.message = self._encode_commit_message(rev, rev.message,
261
def import_commit(self, commit):
262
"""Convert a git commit to a bzr revision.
264
:return: a `bzrlib.revision.Revision` object.
267
raise AssertionError("Commit object can't be None")
268
rev = ForeignRevision(commit.id, self,
269
self.revision_id_foreign_to_bzr(commit.id))
270
rev.parent_ids = tuple([self.revision_id_foreign_to_bzr(p) for p in commit.parents])
271
def decode_using_encoding(rev, commit, encoding):
272
rev.committer = str(commit.committer).decode(encoding)
273
if commit.committer != commit.author:
274
rev.properties['author'] = str(commit.author).decode(encoding)
275
rev.message = self._decode_commit_message(rev, commit.message,
277
if commit.encoding is not None:
278
rev.properties['git-explicit-encoding'] = commit.encoding
279
decode_using_encoding(rev, commit, commit.encoding)
281
for encoding in ('utf-8', 'latin1'):
283
decode_using_encoding(rev, commit, encoding)
284
except UnicodeDecodeError:
287
if encoding != 'utf-8':
288
rev.properties['git-implicit-encoding'] = encoding
290
if commit.commit_time != commit.author_time:
291
rev.properties['author-timestamp'] = str(commit.author_time)
292
if commit.commit_timezone != commit.author_timezone:
293
rev.properties['author-timezone'] = "%d" % commit.author_timezone
294
if commit._author_timezone_neg_utc:
295
rev.properties['author-timezone-neg-utc'] = ""
296
if commit._commit_timezone_neg_utc:
297
rev.properties['commit-timezone-neg-utc'] = ""
298
rev.timestamp = commit.commit_time
299
rev.timezone = commit.commit_timezone
303
class BzrGitMappingv1(BzrGitMapping):
304
revid_prefix = 'git-v1'
308
return self.revid_prefix
311
class BzrGitMappingExperimental(BzrGitMappingv1):
312
revid_prefix = 'git-experimental'
315
def _decode_commit_message(self, rev, message, encoding):
316
message = self._extract_hg_metadata(rev, message)
317
message = self._extract_git_svn_metadata(rev, message)
318
return message.decode(encoding)
320
def _encode_commit_message(self, rev, message, encoding):
321
ret = message.encode(encoding)
322
ret += self._generate_hg_message_tail(rev)
323
ret += self._generate_git_svn_metadata(rev, encoding)
326
def import_commit(self, commit):
327
rev = super(BzrGitMappingExperimental, self).import_commit(commit)
328
rev.properties['converted_revision'] = "git %s\n" % commit.id
332
class GitMappingRegistry(VcsMappingRegistry):
333
"""Registry with available git mappings."""
335
def revision_id_bzr_to_foreign(self, bzr_revid):
336
from dulwich.protocol import ZERO_SHA
337
if bzr_revid == NULL_REVISION:
338
return ZERO_SHA, None
339
if not bzr_revid.startswith("git-"):
340
raise errors.InvalidRevisionId(bzr_revid, None)
341
(mapping_version, git_sha) = bzr_revid.split(":", 1)
342
mapping = self.get(mapping_version)
343
return mapping.revision_id_bzr_to_foreign(bzr_revid)
345
parse_revision_id = revision_id_bzr_to_foreign
348
mapping_registry = GitMappingRegistry()
349
mapping_registry.register_lazy('git-v1', "bzrlib.plugins.git.mapping",
351
mapping_registry.register_lazy('git-experimental',
352
"bzrlib.plugins.git.mapping", "BzrGitMappingExperimental")
353
mapping_registry.set_default('git-v1')
356
class ForeignGit(ForeignVcs):
357
"""The Git Stupid Content Tracker"""
360
def branch_format(self):
361
from bzrlib.plugins.git.branch import GitBranchFormat
362
return GitBranchFormat()
365
def repository_format(self):
366
from bzrlib.plugins.git.repository import GitRepositoryFormat
367
return GitRepositoryFormat()
370
super(ForeignGit, self).__init__(mapping_registry)
371
self.abbreviation = "git"
374
def serialize_foreign_revid(self, foreign_revid):
378
def show_foreign_revid(cls, foreign_revid):
379
return { "git commit": foreign_revid }
382
foreign_git = ForeignGit()
383
default_mapping = mapping_registry.get_default()()
386
def symlink_to_blob(entry):
387
from dulwich.objects import Blob
389
symlink_target = entry.symlink_target
390
if type(symlink_target) == unicode:
391
symlink_target = symlink_target.encode('utf-8')
392
blob.data = symlink_target
396
def mode_is_executable(mode):
397
"""Check if mode should be considered executable."""
398
return bool(mode & 0111)
402
"""Determine the Bazaar inventory kind based on Unix file mode."""
403
entry_kind = (mode & 0700000) / 0100000
406
elif entry_kind == 1:
407
file_kind = (mode & 070000) / 010000
413
return 'tree-reference'
415
raise AssertionError(
416
"Unknown file kind %d, perms=%o." % (file_kind, mode,))
418
raise AssertionError(
419
"Unknown kind, perms=%r." % (mode,))
422
def object_mode(kind, executable):
423
if kind == 'directory':
425
elif kind == 'symlink':
431
mode = stat.S_IFREG | 0644
435
elif kind == 'tree-reference':
436
from dulwich.objects import S_IFGITLINK
442
def entry_mode(entry):
443
"""Determine the git file mode for an inventory entry."""
444
return object_mode(entry.kind, entry.executable)
447
def directory_to_tree(entry, lookup_ie_sha1, unusual_modes):
448
from dulwich.objects import Tree
450
for name, value in entry.children.iteritems():
451
ie = entry.children[name]
453
mode = unusual_modes[ie.file_id]
455
mode = entry_mode(ie)
456
hexsha = lookup_ie_sha1(ie)
457
if hexsha is not None:
458
tree.add(mode, name.encode("utf-8"), hexsha)
459
if entry.parent_id is not None and len(tree) == 0:
460
# Only the root can be an empty tree
465
def extract_unusual_modes(rev):
467
foreign_revid, mapping = mapping_registry.parse_revision_id(
469
except errors.InvalidRevisionId:
472
return mapping.export_unusual_file_modes(rev)
475
def parse_git_svn_id(text):
476
(head, uuid) = text.rsplit(" ", 1)
477
(full_url, rev) = head.rsplit("@", 1)
478
return (full_url, int(rev), uuid)
added
removed
mapping.py
