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.plugins.git.hg import (
48
DEFAULT_FILE_MODE = stat.S_IFREG | 0644
51
def escape_file_id(file_id):
52
return file_id.replace('_', '__').replace(' ', '_s')
55
def unescape_file_id(file_id):
58
while i < len(file_id):
60
ret.append(file_id[i])
62
if file_id[i+1] == '_':
64
elif file_id[i+1] == 's':
67
raise AssertionError("unknown escape character %s" % file_id[i+1])
73
def fix_person_identifier(text):
74
if "<" in text and ">" in text:
76
return "%s <%s>" % (text, text)
79
def warn_escaped(commit, num_escaped):
80
trace.warning("Escaped %d XML-invalid characters in %s. Will be unable "
81
"to regenerate the SHA map.", num_escaped, commit)
84
def warn_unusual_mode(commit, path, mode):
85
trace.mutter("Unusual file mode %o for %s in %s. Storing as revision property. ",
89
def squash_revision(target_repo, rev):
90
"""Remove characters that can't be stored from a revision, if necessary.
92
:param target_repo: Repository in which the revision will be stored
93
:param rev: Revision object, will be modified in-place
95
if not getattr(target_repo._serializer, "squashes_xml_invalid_characters", True):
97
from bzrlib.xml_serializer import escape_invalid_chars
98
rev.message, num_escaped = escape_invalid_chars(rev.message)
100
warn_escaped(rev.foreign_revid, num_escaped)
101
if 'author' in rev.properties:
102
rev.properties['author'], num_escaped = escape_invalid_chars(
103
rev.properties['author'])
105
warn_escaped(rev.foreign_revid, num_escaped)
106
rev.committer, num_escaped = escape_invalid_chars(rev.committer)
108
warn_escaped(rev.foreign_revid, num_escaped)
111
class BzrGitMapping(foreign.VcsMapping):
112
"""Class that maps between Git and Bazaar semantics."""
116
super(BzrGitMapping, self).__init__(foreign_git)
118
def __eq__(self, other):
119
return type(self) == type(other) and self.revid_prefix == other.revid_prefix
122
def revision_id_foreign_to_bzr(cls, git_rev_id):
123
"""Convert a git revision id handle to a Bazaar revision id."""
124
return "%s:%s" % (cls.revid_prefix, git_rev_id)
127
def revision_id_bzr_to_foreign(cls, bzr_rev_id):
128
"""Convert a Bazaar revision id to a git revision id handle."""
129
if not bzr_rev_id.startswith("%s:" % cls.revid_prefix):
130
raise errors.InvalidRevisionId(bzr_rev_id, cls)
131
return bzr_rev_id[len(cls.revid_prefix)+1:], cls()
133
def generate_file_id(self, path):
134
# Git paths are just bytestrings
135
# We must just hope they are valid UTF-8..
138
return escape_file_id(path)
140
def parse_file_id(self, file_id):
141
if file_id == ROOT_ID:
143
return unescape_file_id(file_id)
145
def import_unusual_file_modes(self, rev, unusual_file_modes):
146
if unusual_file_modes:
147
ret = [(name, unusual_file_modes[name]) for name in sorted(unusual_file_modes.keys())]
148
rev.properties['file-modes'] = bencode.bencode(ret)
150
def export_unusual_file_modes(self, rev):
152
return dict([(self.generate_file_id(path), mode) for (path, mode) in bencode.bdecode(rev.properties['file-modes'])])
156
def _generate_hg_message_tail(self, rev):
160
for name in rev.properties:
161
if name == 'hg:extra:branch':
162
branch = rev.properties['hg:extra:branch']
163
elif name.startswith('hg:extra'):
164
extra[name[len('hg:extra:'):]] = base64.b64decode(rev.properties[name])
165
elif name == 'hg:renames':
166
renames = bencode.bdecode(base64.b64decode(rev.properties['hg:renames']))
167
# TODO: Export other properties as 'bzr:' extras?
168
return format_hg_metadata(renames, branch, extra)
170
def _extract_hg_metadata(self, rev, message):
171
(message, renames, branch, extra) = extract_hg_metadata(message)
172
if branch is not None:
173
rev.properties['hg:extra:branch'] = branch
174
for name, value in extra.iteritems():
175
rev.properties['hg:extra:' + name] = base64.b64encode(value)
177
rev.properties['hg:renames'] = base64.b64encode(bencode.bencode([(new, old) for (old, new) in renames.iteritems()]))
180
def _decode_commit_message(self, rev, message):
181
return message.decode("utf-8", "replace")
183
def _encode_commit_message(self, rev, message):
184
return message.encode("utf-8")
186
def export_commit(self, rev, tree_sha, parent_lookup):
187
"""Turn a Bazaar revision in to a Git commit
189
:param tree_sha: Tree sha for the commit
190
:param parent_lookup: Function for looking up the GIT sha equiv of a bzr revision
191
:return dulwich.objects.Commit represent the revision:
193
from dulwich.objects import Commit
195
commit.tree = tree_sha
196
for p in rev.parent_ids:
197
git_p = parent_lookup(p)
198
if git_p is not None:
199
assert len(git_p) == 40, "unexpected length for %r" % git_p
200
commit.parents.append(git_p)
201
commit.committer = fix_person_identifier(rev.committer.encode("utf-8"))
202
commit.author = fix_person_identifier(rev.get_apparent_authors()[0].encode("utf-8"))
203
commit.commit_time = long(rev.timestamp)
204
if 'author-timestamp' in rev.properties:
205
commit.author_time = long(rev.properties['author-timestamp'])
207
commit.author_time = commit.commit_time
208
commit.commit_timezone = rev.timezone
209
if 'author-timezone' in rev.properties:
210
commit.author_timezone = int(rev.properties['author-timezone'])
212
commit.author_timezone = commit.commit_timezone
213
commit.message = self._encode_commit_message(rev, rev.message)
216
def import_commit(self, commit):
217
"""Convert a git commit to a bzr revision.
219
:return: a `bzrlib.revision.Revision` object.
222
raise AssertionError("Commit object can't be None")
223
rev = ForeignRevision(commit.id, self, self.revision_id_foreign_to_bzr(commit.id))
224
rev.parent_ids = tuple([self.revision_id_foreign_to_bzr(p) for p in commit.parents])
225
rev.committer = str(commit.committer).decode("utf-8", "replace")
226
if commit.committer != commit.author:
227
rev.properties['author'] = str(commit.author).decode("utf-8", "replace")
229
if commit.commit_time != commit.author_time:
230
rev.properties['author-timestamp'] = str(commit.author_time)
231
if commit.commit_timezone != commit.author_timezone:
232
rev.properties['author-timezone'] = "%d" % (commit.author_timezone, )
233
rev.timestamp = commit.commit_time
234
rev.timezone = commit.commit_timezone
235
rev.message = self._decode_commit_message(rev, commit.message)
239
class BzrGitMappingv1(BzrGitMapping):
240
revid_prefix = 'git-v1'
244
return self.revid_prefix
247
class BzrGitMappingExperimental(BzrGitMappingv1):
248
revid_prefix = 'git-experimental'
251
def _decode_commit_message(self, rev, message):
252
message = self._extract_hg_metadata(rev, message)
253
return message.decode("utf-8", "replace")
255
def _encode_commit_message(self, rev, message):
256
ret = message.encode("utf-8")
257
ret += self._generate_hg_message_tail(rev)
261
class GitMappingRegistry(VcsMappingRegistry):
262
"""Registry with available git mappings."""
264
def revision_id_bzr_to_foreign(self, bzr_revid):
265
if not bzr_revid.startswith("git-"):
266
raise errors.InvalidRevisionId(bzr_revid, None)
267
(mapping_version, git_sha) = bzr_revid.split(":", 1)
268
mapping = self.get(mapping_version)
269
return mapping.revision_id_bzr_to_foreign(bzr_revid)
271
parse_revision_id = revision_id_bzr_to_foreign
274
mapping_registry = GitMappingRegistry()
275
mapping_registry.register_lazy('git-v1', "bzrlib.plugins.git.mapping",
277
mapping_registry.register_lazy('git-experimental', "bzrlib.plugins.git.mapping",
278
"BzrGitMappingExperimental")
279
mapping_registry.set_default('git-v1')
282
class ForeignGit(ForeignVcs):
283
"""The Git Stupid Content Tracker"""
286
super(ForeignGit, self).__init__(mapping_registry)
289
def show_foreign_revid(cls, foreign_revid):
290
return { "git commit": foreign_revid }
293
foreign_git = ForeignGit()
294
default_mapping = mapping_registry.get_default()()
297
def text_to_blob(texts, entry):
298
from dulwich.objects import Blob
299
text = texts.get_record_stream([(entry.file_id, entry.revision)], 'unordered', True).next().get_bytes_as('fulltext')
305
def symlink_to_blob(entry):
306
from dulwich.objects import Blob
308
blob._text = entry.symlink_target
312
def mode_is_executable(mode):
313
"""Check if mode should be considered executable."""
314
return bool(mode & 0111)
318
"""Determine the Bazaar inventory kind based on Unix file mode."""
319
entry_kind = (mode & 0700000) / 0100000
322
elif entry_kind == 1:
323
file_kind = (mode & 070000) / 010000
329
return 'tree-reference'
331
raise AssertionError(
332
"Unknown file kind %d, perms=%o." % (file_kind, mode,))
334
raise AssertionError(
335
"Unknown kind, perms=%r." % (mode,))
338
def object_mode(kind, executable):
339
if kind == 'directory':
341
elif kind == 'symlink':
344
mode = stat.S_IFREG | 0644
352
def entry_mode(entry):
353
"""Determine the git file mode for an inventory entry."""
354
return object_mode(entry.kind, entry.executable)
357
def directory_to_tree(entry, lookup_ie_sha1, unusual_modes):
358
from dulwich.objects import Tree
360
for name in sorted(entry.children.keys()):
361
ie = entry.children[name]
363
mode = unusual_modes[ie.file_id]
365
mode = entry_mode(ie)
366
hexsha = lookup_ie_sha1(ie)
367
if hexsha is not None:
368
tree.add(mode, name.encode("utf-8"), hexsha)
369
if entry.parent_id is not None and len(tree) == 0:
370
# Only the root can be an empty tree
376
def extract_unusual_modes(rev):
378
foreign_revid, mapping = mapping_registry.parse_revision_id(rev.revision_id)
379
except errors.InvalidRevisionId:
382
return mapping.export_unusual_file_modes(rev)
385
def inventory_to_tree_and_blobs(inventory, texts, mapping, unusual_modes, cur=None):
386
"""Convert a Bazaar tree to a Git tree.
388
:return: Yields tuples with object sha1, object and path
390
from dulwich.objects import Tree
397
# stack contains the set of trees that we haven't
398
# finished constructing
399
for path, entry in inventory.iter_entries():
400
while stack and not path.startswith(osutils.pathjoin(cur, "")):
401
# We've hit a file that's not a child of the previous path
404
yield sha, tree, cur.encode("utf-8")
405
mode = unusual_modes.get(cur.encode("utf-8"), stat.S_IFDIR)
406
t = (mode, urlutils.basename(cur).encode('UTF-8'), sha)
407
cur, tree = stack.pop()
410
if entry.kind == "directory":
411
stack.append((cur, tree))
415
if entry.kind == "file":
416
blob = text_to_blob(texts, entry)
417
elif entry.kind == "symlink":
418
blob = symlink_to_blob(entry)
420
raise AssertionError("Unknown kind %s" % entry.kind)
422
yield sha, blob, path.encode("utf-8")
423
name = urlutils.basename(path).encode("utf-8")
424
mode = unusual_modes.get(path.encode("utf-8"), entry_mode(entry))
425
tree.add(mode, name, sha)
427
while len(stack) > 1:
430
yield sha, tree, cur.encode("utf-8")
431
mode = unusual_modes.get(cur.encode('utf-8'), stat.S_IFDIR)
432
t = (mode, urlutils.basename(cur).encode('UTF-8'), sha)
433
cur, tree = stack.pop()
437
yield tree.id, tree, cur.encode("utf-8")
added
removed
mapping.py
