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
if git_rev_id == "0" * 40:
129
return "%s:%s" % (cls.revid_prefix, git_rev_id)
132
def revision_id_bzr_to_foreign(cls, bzr_rev_id):
133
"""Convert a Bazaar revision id to a git revision id handle."""
134
if not bzr_rev_id.startswith("%s:" % cls.revid_prefix):
135
raise errors.InvalidRevisionId(bzr_rev_id, cls)
136
return bzr_rev_id[len(cls.revid_prefix)+1:], cls()
138
def generate_file_id(self, path):
139
# Git paths are just bytestrings
140
# We must just hope they are valid UTF-8..
143
return escape_file_id(path)
145
def parse_file_id(self, file_id):
146
if file_id == ROOT_ID:
148
return unescape_file_id(file_id)
150
def import_unusual_file_modes(self, rev, unusual_file_modes):
151
if unusual_file_modes:
152
ret = [(path, unusual_file_modes[path])
153
for path in sorted(unusual_file_modes.keys())]
154
rev.properties['file-modes'] = bencode.bencode(ret)
156
def export_unusual_file_modes(self, rev):
158
return dict([(self.generate_file_id(path), mode) for (path, mode) in bencode.bdecode(rev.properties['file-modes'].encode("utf-8"))])
162
def _generate_git_svn_metadata(self, rev, encoding):
164
return "\ngit-svn-id: %s\n" % rev.properties["git-svn-id"].encode(
169
def _generate_hg_message_tail(self, rev):
173
for name in rev.properties:
174
if name == 'hg:extra:branch':
175
branch = rev.properties['hg:extra:branch']
176
elif name.startswith('hg:extra'):
177
extra[name[len('hg:extra:'):]] = base64.b64decode(
178
rev.properties[name])
179
elif name == 'hg:renames':
180
renames = bencode.bdecode(base64.b64decode(
181
rev.properties['hg:renames']))
182
# TODO: Export other properties as 'bzr:' extras?
183
ret = format_hg_metadata(renames, branch, extra)
184
assert isinstance(ret, str)
187
def _extract_git_svn_metadata(self, rev, message):
188
lines = message.split("\n")
189
if not (lines[-1] == "" and lines[-2].startswith("git-svn-id:")):
191
git_svn_id = lines[-2].split(": ", 1)[1]
192
rev.properties['git-svn-id'] = git_svn_id
193
(url, rev, uuid) = parse_git_svn_id(git_svn_id)
194
# FIXME: Convert this to converted-from property somehow..
195
ret = "\n".join(lines[:-2])
196
assert isinstance(ret, str)
199
def _extract_hg_metadata(self, rev, message):
200
(message, renames, branch, extra) = extract_hg_metadata(message)
201
if branch is not None:
202
rev.properties['hg:extra:branch'] = branch
203
for name, value in extra.iteritems():
204
rev.properties['hg:extra:' + name] = base64.b64encode(value)
206
rev.properties['hg:renames'] = base64.b64encode(bencode.bencode(
207
[(new, old) for (old, new) in renames.iteritems()]))
210
def _decode_commit_message(self, rev, message, encoding):
211
return message.decode(encoding)
213
def _encode_commit_message(self, rev, message, encoding):
214
return message.encode(encoding)
216
def export_commit(self, rev, tree_sha, parent_lookup):
217
"""Turn a Bazaar revision in to a Git commit
219
:param tree_sha: Tree sha for the commit
220
:param parent_lookup: Function for looking up the GIT sha equiv of a
222
:return dulwich.objects.Commit represent the revision:
224
from dulwich.objects import Commit
226
commit.tree = tree_sha
227
for p in rev.parent_ids:
229
git_p = parent_lookup(p)
232
if git_p is not None:
233
assert len(git_p) == 40, "unexpected length for %r" % git_p
234
commit.parents.append(git_p)
236
encoding = rev.properties['git-explicit-encoding']
238
encoding = rev.properties.get('git-implicit-encoding', 'utf-8')
239
commit.encoding = rev.properties.get('git-explicit-encoding')
240
commit.committer = fix_person_identifier(rev.committer.encode(
242
commit.author = fix_person_identifier(
243
rev.get_apparent_authors()[0].encode(encoding))
244
commit.commit_time = long(rev.timestamp)
245
if 'author-timestamp' in rev.properties:
246
commit.author_time = long(rev.properties['author-timestamp'])
248
commit.author_time = commit.commit_time
249
commit._commit_timezone_neg_utc = "commit-timezone-neg-utc" in rev.properties
250
commit.commit_timezone = rev.timezone
251
commit._author_timezone_neg_utc = "author-timezone-neg-utc" in rev.properties
252
if 'author-timezone' in rev.properties:
253
commit.author_timezone = int(rev.properties['author-timezone'])
255
commit.author_timezone = commit.commit_timezone
256
commit.message = self._encode_commit_message(rev, rev.message,
260
def import_commit(self, commit):
261
"""Convert a git commit to a bzr revision.
263
:return: a `bzrlib.revision.Revision` object.
266
raise AssertionError("Commit object can't be None")
267
rev = ForeignRevision(commit.id, self,
268
self.revision_id_foreign_to_bzr(commit.id))
269
rev.parent_ids = tuple([self.revision_id_foreign_to_bzr(p) for p in commit.parents])
270
def decode_using_encoding(rev, commit, encoding):
271
rev.committer = str(commit.committer).decode(encoding)
272
if commit.committer != commit.author:
273
rev.properties['author'] = str(commit.author).decode(encoding)
274
rev.message = self._decode_commit_message(rev, commit.message,
276
if commit.encoding is not None:
277
rev.properties['git-explicit-encoding'] = commit.encoding
278
decode_using_encoding(rev, commit, commit.encoding)
280
for encoding in ('utf-8', 'latin1'):
282
decode_using_encoding(rev, commit, encoding)
283
except UnicodeDecodeError:
286
if encoding != 'utf-8':
287
rev.properties['git-implicit-encoding'] = encoding
289
if commit.commit_time != commit.author_time:
290
rev.properties['author-timestamp'] = str(commit.author_time)
291
if commit.commit_timezone != commit.author_timezone:
292
rev.properties['author-timezone'] = "%d" % commit.author_timezone
293
if commit._author_timezone_neg_utc:
294
rev.properties['author-timezone-neg-utc'] = ""
295
if commit._commit_timezone_neg_utc:
296
rev.properties['commit-timezone-neg-utc'] = ""
297
rev.timestamp = commit.commit_time
298
rev.timezone = commit.commit_timezone
302
class BzrGitMappingv1(BzrGitMapping):
303
revid_prefix = 'git-v1'
307
return self.revid_prefix
310
class BzrGitMappingExperimental(BzrGitMappingv1):
311
revid_prefix = 'git-experimental'
314
def _decode_commit_message(self, rev, message, encoding):
315
message = self._extract_hg_metadata(rev, message)
316
message = self._extract_git_svn_metadata(rev, message)
317
return message.decode(encoding)
319
def _encode_commit_message(self, rev, message, encoding):
320
ret = message.encode(encoding)
321
ret += self._generate_hg_message_tail(rev)
322
ret += self._generate_git_svn_metadata(rev, encoding)
325
def import_commit(self, commit):
326
rev = super(BzrGitMappingExperimental, self).import_commit(commit)
327
rev.properties['converted_revision'] = "git %s\n" % commit.id
331
class GitMappingRegistry(VcsMappingRegistry):
332
"""Registry with available git mappings."""
334
def revision_id_bzr_to_foreign(self, bzr_revid):
335
if bzr_revid == NULL_REVISION:
336
return "0" * 20, None
337
if not bzr_revid.startswith("git-"):
338
raise errors.InvalidRevisionId(bzr_revid, None)
339
(mapping_version, git_sha) = bzr_revid.split(":", 1)
340
mapping = self.get(mapping_version)
341
return mapping.revision_id_bzr_to_foreign(bzr_revid)
343
parse_revision_id = revision_id_bzr_to_foreign
346
mapping_registry = GitMappingRegistry()
347
mapping_registry.register_lazy('git-v1', "bzrlib.plugins.git.mapping",
349
mapping_registry.register_lazy('git-experimental',
350
"bzrlib.plugins.git.mapping", "BzrGitMappingExperimental")
351
mapping_registry.set_default('git-v1')
354
class ForeignGit(ForeignVcs):
355
"""The Git Stupid Content Tracker"""
358
def branch_format(self):
359
from bzrlib.plugins.git.branch import GitBranchFormat
360
return GitBranchFormat()
363
def repository_format(self):
364
from bzrlib.plugins.git.repository import GitRepositoryFormat
365
return GitRepositoryFormat()
368
super(ForeignGit, self).__init__(mapping_registry)
369
self.abbreviation = "git"
372
def serialize_foreign_revid(self, foreign_revid):
376
def show_foreign_revid(cls, foreign_revid):
377
return { "git commit": foreign_revid }
380
foreign_git = ForeignGit()
381
default_mapping = mapping_registry.get_default()()
384
def symlink_to_blob(entry):
385
from dulwich.objects import Blob
387
symlink_target = entry.symlink_target
388
if type(symlink_target) == unicode:
389
symlink_target = symlink_target.encode('utf-8')
390
blob.data = symlink_target
394
def mode_is_executable(mode):
395
"""Check if mode should be considered executable."""
396
return bool(mode & 0111)
400
"""Determine the Bazaar inventory kind based on Unix file mode."""
401
entry_kind = (mode & 0700000) / 0100000
404
elif entry_kind == 1:
405
file_kind = (mode & 070000) / 010000
411
return 'tree-reference'
413
raise AssertionError(
414
"Unknown file kind %d, perms=%o." % (file_kind, mode,))
416
raise AssertionError(
417
"Unknown kind, perms=%r." % (mode,))
420
def object_mode(kind, executable):
421
if kind == 'directory':
423
elif kind == 'symlink':
429
mode = stat.S_IFREG | 0644
433
elif kind == 'tree-reference':
434
from dulwich.objects import S_IFGITLINK
440
def entry_mode(entry):
441
"""Determine the git file mode for an inventory entry."""
442
return object_mode(entry.kind, entry.executable)
445
def directory_to_tree(entry, lookup_ie_sha1, unusual_modes):
446
from dulwich.objects import Tree
448
for name, value in entry.children.iteritems():
449
ie = entry.children[name]
451
mode = unusual_modes[ie.file_id]
453
mode = entry_mode(ie)
454
hexsha = lookup_ie_sha1(ie)
455
if hexsha is not None:
456
tree.add(mode, name.encode("utf-8"), hexsha)
457
if entry.parent_id is not None and len(tree) == 0:
458
# Only the root can be an empty tree
463
def extract_unusual_modes(rev):
465
foreign_revid, mapping = mapping_registry.parse_revision_id(
467
except errors.InvalidRevisionId:
470
return mapping.export_unusual_file_modes(rev)
473
def parse_git_svn_id(text):
474
(head, uuid) = text.rsplit(" ", 1)
475
(full_url, rev) = head.rsplit("@", 1)
476
return (full_url, int(rev), uuid)
added
removed
mapping.py
