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."""
32
from bzrlib import bencode
34
from bzrlib.util import bencode
35
from bzrlib.inventory import (
38
from bzrlib.foreign import (
43
from bzrlib.revision import (
46
from bzrlib.plugins.git.hg import (
51
DEFAULT_FILE_MODE = stat.S_IFREG | 0644
54
def escape_file_id(file_id):
55
return file_id.replace('_', '__').replace(' ', '_s')
58
def unescape_file_id(file_id):
61
while i < len(file_id):
63
ret.append(file_id[i])
65
if file_id[i+1] == '_':
67
elif file_id[i+1] == 's':
70
raise AssertionError("unknown escape character %s" %
77
def fix_person_identifier(text):
78
if "<" in text and ">" in text:
80
return "%s <%s>" % (text, text)
83
def warn_escaped(commit, num_escaped):
84
trace.warning("Escaped %d XML-invalid characters in %s. Will be unable "
85
"to regenerate the SHA map.", num_escaped, commit)
88
def warn_unusual_mode(commit, path, mode):
89
trace.mutter("Unusual file mode %o for %s in %s. Storing as revision "
90
"property. ", mode, path, commit)
93
def squash_revision(target_repo, rev):
94
"""Remove characters that can't be stored from a revision, if necessary.
96
:param target_repo: Repository in which the revision will be stored
97
:param rev: Revision object, will be modified in-place
99
if not getattr(target_repo._serializer, "squashes_xml_invalid_characters", True):
101
from bzrlib.xml_serializer import escape_invalid_chars
102
rev.message, num_escaped = escape_invalid_chars(rev.message)
104
warn_escaped(rev.foreign_revid, num_escaped)
105
if 'author' in rev.properties:
106
rev.properties['author'], num_escaped = escape_invalid_chars(
107
rev.properties['author'])
109
warn_escaped(rev.foreign_revid, num_escaped)
110
rev.committer, num_escaped = escape_invalid_chars(rev.committer)
112
warn_escaped(rev.foreign_revid, num_escaped)
115
class BzrGitMapping(foreign.VcsMapping):
116
"""Class that maps between Git and Bazaar semantics."""
120
super(BzrGitMapping, self).__init__(foreign_git)
122
def __eq__(self, other):
123
return (type(self) == type(other) and
124
self.revid_prefix == other.revid_prefix)
127
def revision_id_foreign_to_bzr(cls, git_rev_id):
128
"""Convert a git revision id handle to a Bazaar revision id."""
129
if git_rev_id == "0" * 40:
131
return "%s:%s" % (cls.revid_prefix, git_rev_id)
134
def revision_id_bzr_to_foreign(cls, bzr_rev_id):
135
"""Convert a Bazaar revision id to a git revision id handle."""
136
if not bzr_rev_id.startswith("%s:" % cls.revid_prefix):
137
raise errors.InvalidRevisionId(bzr_rev_id, cls)
138
return bzr_rev_id[len(cls.revid_prefix)+1:], cls()
140
def generate_file_id(self, path):
141
# Git paths are just bytestrings
142
# We must just hope they are valid UTF-8..
145
return escape_file_id(path)
147
def parse_file_id(self, file_id):
148
if file_id == ROOT_ID:
150
return unescape_file_id(file_id)
152
def import_unusual_file_modes(self, rev, unusual_file_modes):
153
if unusual_file_modes:
154
ret = [(name, unusual_file_modes[name])
155
for name in sorted(unusual_file_modes.keys())]
156
rev.properties['file-modes'] = bencode.bencode(ret)
158
def export_unusual_file_modes(self, rev):
160
return dict([(self.generate_file_id(path), mode) for (path, mode) in bencode.bdecode(rev.properties['file-modes'].encode("utf-8"))])
164
def _generate_git_svn_metadata(self, rev, encoding):
166
return "\ngit-svn-id: %s\n" % rev.properties["git-svn-id"].encode(
171
def _generate_hg_message_tail(self, rev):
175
for name in rev.properties:
176
if name == 'hg:extra:branch':
177
branch = rev.properties['hg:extra:branch']
178
elif name.startswith('hg:extra'):
179
extra[name[len('hg:extra:'):]] = base64.b64decode(
180
rev.properties[name])
181
elif name == 'hg:renames':
182
renames = bencode.bdecode(base64.b64decode(
183
rev.properties['hg:renames']))
184
# TODO: Export other properties as 'bzr:' extras?
185
ret = format_hg_metadata(renames, branch, extra)
186
assert isinstance(ret, str)
189
def _extract_git_svn_metadata(self, rev, message):
190
lines = message.split("\n")
191
if not (lines[-1] == "" and lines[-2].startswith("git-svn-id:")):
193
git_svn_id = lines[-2].split(": ", 1)[1]
194
rev.properties['git-svn-id'] = git_svn_id
195
(url, rev, uuid) = parse_git_svn_id(git_svn_id)
196
# FIXME: Convert this to converted-from property somehow..
197
ret = "\n".join(lines[:-2])
198
assert isinstance(ret, str)
201
def _extract_hg_metadata(self, rev, message):
202
(message, renames, branch, extra) = extract_hg_metadata(message)
203
if branch is not None:
204
rev.properties['hg:extra:branch'] = branch
205
for name, value in extra.iteritems():
206
rev.properties['hg:extra:' + name] = base64.b64encode(value)
208
rev.properties['hg:renames'] = base64.b64encode(bencode.bencode(
209
[(new, old) for (old, new) in renames.iteritems()]))
212
def _decode_commit_message(self, rev, message, encoding):
213
return message.decode(encoding)
215
def _encode_commit_message(self, rev, message, encoding):
216
return message.encode(encoding)
218
def export_commit(self, rev, tree_sha, parent_lookup):
219
"""Turn a Bazaar revision in to a Git commit
221
:param tree_sha: Tree sha for the commit
222
:param parent_lookup: Function for looking up the GIT sha equiv of a
224
:return dulwich.objects.Commit represent the revision:
226
from dulwich.objects import Commit
228
commit.tree = tree_sha
229
for p in rev.parent_ids:
231
git_p = parent_lookup(p)
234
if git_p is not None:
235
assert len(git_p) == 40, "unexpected length for %r" % git_p
236
commit.parents.append(git_p)
238
encoding = rev.properties['git-explicit-encoding']
240
encoding = rev.properties.get('git-implicit-encoding', 'utf-8')
241
commit.encoding = rev.properties.get('git-explicit-encoding')
242
commit.committer = fix_person_identifier(rev.committer.encode(
244
commit.author = fix_person_identifier(
245
rev.get_apparent_authors()[0].encode(encoding))
246
commit.commit_time = long(rev.timestamp)
247
if 'author-timestamp' in rev.properties:
248
commit.author_time = long(rev.properties['author-timestamp'])
250
commit.author_time = commit.commit_time
251
commit.commit_timezone = rev.timezone
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
rev.timestamp = commit.commit_time
294
rev.timezone = commit.commit_timezone
298
class BzrGitMappingv1(BzrGitMapping):
299
revid_prefix = 'git-v1'
303
return self.revid_prefix
306
class BzrGitMappingExperimental(BzrGitMappingv1):
307
revid_prefix = 'git-experimental'
310
def _decode_commit_message(self, rev, message, encoding):
311
message = self._extract_hg_metadata(rev, message)
312
message = self._extract_git_svn_metadata(rev, message)
313
return message.decode(encoding)
315
def _encode_commit_message(self, rev, message, encoding):
316
ret = message.encode(encoding)
317
ret += self._generate_hg_message_tail(rev)
318
ret += self._generate_git_svn_metadata(rev, encoding)
321
def import_commit(self, commit):
322
rev = super(BzrGitMappingExperimental, self).import_commit(commit)
323
rev.properties['converted_revision'] = "git %s\n" % commit.id
327
class GitMappingRegistry(VcsMappingRegistry):
328
"""Registry with available git mappings."""
330
def revision_id_bzr_to_foreign(self, bzr_revid):
331
if bzr_revid == NULL_REVISION:
332
return "0" * 20, None
333
if not bzr_revid.startswith("git-"):
334
raise errors.InvalidRevisionId(bzr_revid, None)
335
(mapping_version, git_sha) = bzr_revid.split(":", 1)
336
mapping = self.get(mapping_version)
337
return mapping.revision_id_bzr_to_foreign(bzr_revid)
339
parse_revision_id = revision_id_bzr_to_foreign
342
mapping_registry = GitMappingRegistry()
343
mapping_registry.register_lazy('git-v1', "bzrlib.plugins.git.mapping",
345
mapping_registry.register_lazy('git-experimental',
346
"bzrlib.plugins.git.mapping", "BzrGitMappingExperimental")
347
mapping_registry.set_default('git-v1')
350
class ForeignGit(ForeignVcs):
351
"""The Git Stupid Content Tracker"""
354
def branch_format(self):
355
from bzrlib.plugins.git.branch import GitBranchFormat
356
return GitBranchFormat()
359
def repository_format(self):
360
from bzrlib.plugins.git.repository import GitRepositoryFormat
361
return GitRepositoryFormat()
364
super(ForeignGit, self).__init__(mapping_registry)
365
self.abbreviation = "git"
368
def serialize_foreign_revid(self, foreign_revid):
372
def show_foreign_revid(cls, foreign_revid):
373
return { "git commit": foreign_revid }
376
foreign_git = ForeignGit()
377
default_mapping = mapping_registry.get_default()()
380
def symlink_to_blob(entry):
381
from dulwich.objects import Blob
383
symlink_target = entry.symlink_target
384
if type(symlink_target) == unicode:
385
symlink_target = symlink_target.encode('utf-8')
386
blob.data = symlink_target
390
def mode_is_executable(mode):
391
"""Check if mode should be considered executable."""
392
return bool(mode & 0111)
396
"""Determine the Bazaar inventory kind based on Unix file mode."""
397
entry_kind = (mode & 0700000) / 0100000
400
elif entry_kind == 1:
401
file_kind = (mode & 070000) / 010000
407
return 'tree-reference'
409
raise AssertionError(
410
"Unknown file kind %d, perms=%o." % (file_kind, mode,))
412
raise AssertionError(
413
"Unknown kind, perms=%r." % (mode,))
416
def object_mode(kind, executable):
417
if kind == 'directory':
419
elif kind == 'symlink':
425
mode = stat.S_IFREG | 0644
429
elif kind == 'tree-reference':
430
from dulwich.objects import S_IFGITLINK
436
def entry_mode(entry):
437
"""Determine the git file mode for an inventory entry."""
438
return object_mode(entry.kind, entry.executable)
441
def directory_to_tree(entry, lookup_ie_sha1, unusual_modes):
442
from dulwich.objects import Tree
444
for name, value in entry.children.iteritems():
445
ie = entry.children[name]
447
mode = unusual_modes[ie.file_id]
449
mode = entry_mode(ie)
450
hexsha = lookup_ie_sha1(ie)
451
if hexsha is not None:
452
tree.add(mode, name.encode("utf-8"), hexsha)
453
if entry.parent_id is not None and len(tree) == 0:
454
# Only the root can be an empty tree
459
def extract_unusual_modes(rev):
461
foreign_revid, mapping = mapping_registry.parse_revision_id(
463
except errors.InvalidRevisionId:
466
return mapping.export_unusual_file_modes(rev)
469
def inventory_to_tree_and_blobs(inventory, texts, mapping, unusual_modes,
471
"""Convert a Bazaar tree to a Git tree.
473
:return: Yields tuples with object sha1, object and path
475
from dulwich.objects import Tree
482
# stack contains the set of trees that we haven't
483
# finished constructing
484
for path, entry in inventory.iter_entries():
485
while stack and not path.startswith(osutils.pathjoin(cur, "")):
486
# We've hit a file that's not a child of the previous path
488
yield sha, tree, cur.encode("utf-8")
489
mode = unusual_modes.get(cur.encode("utf-8"), stat.S_IFDIR)
490
t = (mode, urlutils.basename(cur).encode('UTF-8'), sha)
491
cur, tree = stack.pop()
494
if entry.kind == "directory":
495
stack.append((cur, tree))
499
if entry.kind == "file":
500
from dulwich.objects import Blob
501
stream = texts.get_record_stream(
502
[(entry.file_id, entry.revision)], 'unordered', True)
503
text = stream.next().get_bytes_as('fulltext')
506
elif entry.kind == "symlink":
507
blob = symlink_to_blob(entry)
509
raise AssertionError("Unknown kind %s" % entry.kind)
511
yield sha, blob, path.encode("utf-8")
512
name = urlutils.basename(path).encode("utf-8")
513
mode = unusual_modes.get(path.encode("utf-8"), entry_mode(entry))
514
tree.add(mode, name, sha)
516
while len(stack) > 1:
518
yield sha, tree, cur.encode("utf-8")
519
mode = unusual_modes.get(cur.encode('utf-8'), stat.S_IFDIR)
520
t = (mode, urlutils.basename(cur).encode('UTF-8'), sha)
521
cur, tree = stack.pop()
524
yield tree.id, tree, cur.encode("utf-8")
527
def parse_git_svn_id(text):
528
(head, uuid) = text.rsplit(" ", 1)
529
(full_url, rev) = head.rsplit("@", 1)
530
return (full_url, int(rev), uuid)
added
removed
mapping.py
