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_git_svn_metadata(self, rev):
158
return "\ngit-svn-id: %s\n" % rev.properties["git-svn-id"].encode("utf-8")
162
def _generate_hg_message_tail(self, rev):
166
for name in rev.properties:
167
if name == 'hg:extra:branch':
168
branch = rev.properties['hg:extra:branch']
169
elif name.startswith('hg:extra'):
170
extra[name[len('hg:extra:'):]] = base64.b64decode(rev.properties[name])
171
elif name == 'hg:renames':
172
renames = bencode.bdecode(base64.b64decode(rev.properties['hg:renames']))
173
# TODO: Export other properties as 'bzr:' extras?
174
ret = format_hg_metadata(renames, branch, extra)
175
assert isinstance(ret, str)
178
def _extract_git_svn_metadata(self, rev, message):
179
lines = message.split("\n")
180
if not (lines[-1] == "" and lines[-2].startswith("git-svn-id:")):
182
git_svn_id = lines[-2].split(": ", 1)[1]
183
rev.properties['git-svn-id'] = git_svn_id
184
(url, rev, uuid) = parse_git_svn_id(git_svn_id)
185
# FIXME: Convert this to converted-from property somehow..
186
ret = "\n".join(lines[:-2])
187
assert isinstance(ret, str)
190
def _extract_hg_metadata(self, rev, message):
191
(message, renames, branch, extra) = extract_hg_metadata(message)
192
if branch is not None:
193
rev.properties['hg:extra:branch'] = branch
194
for name, value in extra.iteritems():
195
rev.properties['hg:extra:' + name] = base64.b64encode(value)
197
rev.properties['hg:renames'] = base64.b64encode(bencode.bencode([(new, old) for (old, new) in renames.iteritems()]))
200
def _decode_commit_message(self, rev, message):
201
return message.decode("utf-8", "replace")
203
def _encode_commit_message(self, rev, message):
204
return message.encode("utf-8")
206
def export_commit(self, rev, tree_sha, parent_lookup):
207
"""Turn a Bazaar revision in to a Git commit
209
:param tree_sha: Tree sha for the commit
210
:param parent_lookup: Function for looking up the GIT sha equiv of a bzr revision
211
:return dulwich.objects.Commit represent the revision:
213
from dulwich.objects import Commit
215
commit.tree = tree_sha
216
for p in rev.parent_ids:
217
git_p = parent_lookup(p)
218
if git_p is not None:
219
assert len(git_p) == 40, "unexpected length for %r" % git_p
220
commit.parents.append(git_p)
221
commit.committer = fix_person_identifier(rev.committer.encode("utf-8"))
222
commit.author = fix_person_identifier(rev.get_apparent_authors()[0].encode("utf-8"))
223
commit.commit_time = long(rev.timestamp)
224
if 'author-timestamp' in rev.properties:
225
commit.author_time = long(rev.properties['author-timestamp'])
227
commit.author_time = commit.commit_time
228
commit.commit_timezone = rev.timezone
229
if 'author-timezone' in rev.properties:
230
commit.author_timezone = int(rev.properties['author-timezone'])
232
commit.author_timezone = commit.commit_timezone
233
commit.message = self._encode_commit_message(rev, rev.message)
236
def import_commit(self, commit):
237
"""Convert a git commit to a bzr revision.
239
:return: a `bzrlib.revision.Revision` object.
242
raise AssertionError("Commit object can't be None")
243
rev = ForeignRevision(commit.id, self, self.revision_id_foreign_to_bzr(commit.id))
244
rev.parent_ids = tuple([self.revision_id_foreign_to_bzr(p) for p in commit.parents])
245
rev.committer = str(commit.committer).decode("utf-8", "replace")
246
if commit.committer != commit.author:
247
rev.properties['author'] = str(commit.author).decode("utf-8", "replace")
249
if commit.commit_time != commit.author_time:
250
rev.properties['author-timestamp'] = str(commit.author_time)
251
if commit.commit_timezone != commit.author_timezone:
252
rev.properties['author-timezone'] = "%d" % (commit.author_timezone, )
253
rev.timestamp = commit.commit_time
254
rev.timezone = commit.commit_timezone
255
rev.message = self._decode_commit_message(rev, commit.message)
259
class BzrGitMappingv1(BzrGitMapping):
260
revid_prefix = 'git-v1'
264
return self.revid_prefix
267
class BzrGitMappingExperimental(BzrGitMappingv1):
268
revid_prefix = 'git-experimental'
271
def _decode_commit_message(self, rev, message):
272
message = self._extract_hg_metadata(rev, message)
273
message = self._extract_git_svn_metadata(rev, message)
274
return message.decode("utf-8", "replace")
276
def _encode_commit_message(self, rev, message):
277
ret = message.encode("utf-8")
278
ret += self._generate_hg_message_tail(rev)
279
ret += self._generate_git_svn_metadata(rev)
282
def import_commit(self, commit):
283
rev = super(BzrGitMappingExperimental, self).import_commit(commit)
284
rev.properties['converted_revision'] = "git %s\n" % commit.id
288
class GitMappingRegistry(VcsMappingRegistry):
289
"""Registry with available git mappings."""
291
def revision_id_bzr_to_foreign(self, bzr_revid):
292
if not bzr_revid.startswith("git-"):
293
raise errors.InvalidRevisionId(bzr_revid, None)
294
(mapping_version, git_sha) = bzr_revid.split(":", 1)
295
mapping = self.get(mapping_version)
296
return mapping.revision_id_bzr_to_foreign(bzr_revid)
298
parse_revision_id = revision_id_bzr_to_foreign
301
mapping_registry = GitMappingRegistry()
302
mapping_registry.register_lazy('git-v1', "bzrlib.plugins.git.mapping",
304
mapping_registry.register_lazy('git-experimental', "bzrlib.plugins.git.mapping",
305
"BzrGitMappingExperimental")
306
mapping_registry.set_default('git-v1')
309
class ForeignGit(ForeignVcs):
310
"""The Git Stupid Content Tracker"""
313
def branch_format(self):
314
from bzrlib.plugins.git.branch import GitBranchFormat
315
return GitBranchFormat()
318
def repository_format(self):
319
from bzrlib.plugins.git.repository import GitRepositoryFormat
320
return GitRepositoryFormat()
323
super(ForeignGit, self).__init__(mapping_registry)
324
self.abbreviation = "git"
327
def serialize_foreign_revid(self, foreign_revid):
331
def show_foreign_revid(cls, foreign_revid):
332
return { "git commit": foreign_revid }
335
foreign_git = ForeignGit()
336
default_mapping = mapping_registry.get_default()()
339
def text_to_blob(texts, entry):
340
from dulwich.objects import Blob
341
text = texts.get_record_stream([(entry.file_id, entry.revision)], 'unordered', True).next().get_bytes_as('fulltext')
347
def symlink_to_blob(entry):
348
from dulwich.objects import Blob
350
blob._text = entry.symlink_target
354
def mode_is_executable(mode):
355
"""Check if mode should be considered executable."""
356
return bool(mode & 0111)
360
"""Determine the Bazaar inventory kind based on Unix file mode."""
361
entry_kind = (mode & 0700000) / 0100000
364
elif entry_kind == 1:
365
file_kind = (mode & 070000) / 010000
371
return 'tree-reference'
373
raise AssertionError(
374
"Unknown file kind %d, perms=%o." % (file_kind, mode,))
376
raise AssertionError(
377
"Unknown kind, perms=%r." % (mode,))
380
def object_mode(kind, executable):
381
if kind == 'directory':
383
elif kind == 'symlink':
386
mode = stat.S_IFREG | 0644
394
def entry_mode(entry):
395
"""Determine the git file mode for an inventory entry."""
396
return object_mode(entry.kind, entry.executable)
399
def directory_to_tree(entry, lookup_ie_sha1, unusual_modes):
400
from dulwich.objects import Tree
402
for name in sorted(entry.children.keys()):
403
ie = entry.children[name]
405
mode = unusual_modes[ie.file_id]
407
mode = entry_mode(ie)
408
hexsha = lookup_ie_sha1(ie)
409
if hexsha is not None:
410
tree.add(mode, name.encode("utf-8"), hexsha)
411
if entry.parent_id is not None and len(tree) == 0:
412
# Only the root can be an empty tree
418
def extract_unusual_modes(rev):
420
foreign_revid, mapping = mapping_registry.parse_revision_id(rev.revision_id)
421
except errors.InvalidRevisionId:
424
return mapping.export_unusual_file_modes(rev)
427
def inventory_to_tree_and_blobs(inventory, texts, mapping, unusual_modes, cur=None):
428
"""Convert a Bazaar tree to a Git tree.
430
:return: Yields tuples with object sha1, object and path
432
from dulwich.objects import Tree
439
# stack contains the set of trees that we haven't
440
# finished constructing
441
for path, entry in inventory.iter_entries():
442
while stack and not path.startswith(osutils.pathjoin(cur, "")):
443
# We've hit a file that's not a child of the previous path
446
yield sha, tree, cur.encode("utf-8")
447
mode = unusual_modes.get(cur.encode("utf-8"), stat.S_IFDIR)
448
t = (mode, urlutils.basename(cur).encode('UTF-8'), sha)
449
cur, tree = stack.pop()
452
if entry.kind == "directory":
453
stack.append((cur, tree))
457
if entry.kind == "file":
458
blob = text_to_blob(texts, entry)
459
elif entry.kind == "symlink":
460
blob = symlink_to_blob(entry)
462
raise AssertionError("Unknown kind %s" % entry.kind)
464
yield sha, blob, path.encode("utf-8")
465
name = urlutils.basename(path).encode("utf-8")
466
mode = unusual_modes.get(path.encode("utf-8"), entry_mode(entry))
467
tree.add(mode, name, sha)
469
while len(stack) > 1:
472
yield sha, tree, cur.encode("utf-8")
473
mode = unusual_modes.get(cur.encode('utf-8'), stat.S_IFDIR)
474
t = (mode, urlutils.basename(cur).encode('UTF-8'), sha)
475
cur, tree = stack.pop()
479
yield tree.id, tree, cur.encode("utf-8")
482
def parse_git_svn_id(text):
483
(head, uuid) = text.rsplit(" ", 1)
484
(full_url, rev) = head.rsplit("@", 1)
485
return (full_url, int(rev), uuid)
added
removed
mapping.py
