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."""
31
from bzrlib import bencode
33
from bzrlib.util import bencode
34
from bzrlib.inventory import (
37
from bzrlib.foreign import (
43
DEFAULT_FILE_MODE = stat.S_IFREG | 0644
46
def escape_file_id(file_id):
47
return file_id.replace('_', '__').replace(' ', '_s')
50
def unescape_file_id(file_id):
53
while i < len(file_id):
55
ret.append(file_id[i])
57
if file_id[i+1] == '_':
59
elif file_id[i+1] == 's':
62
raise AssertionError("unknown escape character %s" % file_id[i+1])
68
def fix_person_identifier(text):
69
if "<" in text and ">" in text:
71
return "%s <%s>" % (text, text)
74
def warn_escaped(commit, num_escaped):
75
trace.warning("Escaped %d XML-invalid characters in %s. Will be unable "
76
"to regenerate the SHA map.", num_escaped, commit)
79
def warn_unusual_mode(commit, path, mode):
80
trace.mutter("Unusual file mode %o for %s in %s. Storing as revision property. ",
84
def squash_revision(target_repo, rev):
85
"""Remove characters that can't be stored from a revision, if necessary.
87
:param target_repo: Repository in which the revision will be stored
88
:param rev: Revision object, will be modified in-place
90
if not getattr(target_repo._serializer, "squashes_xml_invalid_characters", True):
92
from bzrlib.xml_serializer import escape_invalid_chars
93
rev.message, num_escaped = escape_invalid_chars(rev.message)
95
warn_escaped(rev.foreign_revid, num_escaped)
96
if 'author' in rev.properties:
97
rev.properties['author'], num_escaped = escape_invalid_chars(
98
rev.properties['author'])
100
warn_escaped(rev.foreign_revid, num_escaped)
101
rev.committer, num_escaped = escape_invalid_chars(rev.committer)
103
warn_escaped(rev.foreign_revid, num_escaped)
106
class BzrGitMapping(foreign.VcsMapping):
107
"""Class that maps between Git and Bazaar semantics."""
111
super(BzrGitMapping, self).__init__(foreign_git)
113
def __eq__(self, other):
114
return type(self) == type(other) and self.revid_prefix == other.revid_prefix
117
def revision_id_foreign_to_bzr(cls, git_rev_id):
118
"""Convert a git revision id handle to a Bazaar revision id."""
119
return "%s:%s" % (cls.revid_prefix, git_rev_id)
122
def revision_id_bzr_to_foreign(cls, bzr_rev_id):
123
"""Convert a Bazaar revision id to a git revision id handle."""
124
if not bzr_rev_id.startswith("%s:" % cls.revid_prefix):
125
raise errors.InvalidRevisionId(bzr_rev_id, cls)
126
return bzr_rev_id[len(cls.revid_prefix)+1:], cls()
128
def generate_file_id(self, path):
129
# Git paths are just bytestrings
130
# We must just hope they are valid UTF-8..
133
return escape_file_id(path)
135
def parse_file_id(self, file_id):
136
if file_id == ROOT_ID:
138
return unescape_file_id(file_id)
140
def import_unusual_file_modes(self, rev, unusual_file_modes):
141
if unusual_file_modes:
142
ret = [(name, unusual_file_modes[name]) for name in sorted(unusual_file_modes.keys())]
143
rev.properties['file-modes'] = bencode.bencode(ret)
145
def export_unusual_file_modes(self, rev):
147
return dict([(self.generate_file_id(path), mode) for (path, mode) in bencode.bdecode(rev.properties['file-modes'])])
151
def import_commit(self, commit):
152
"""Convert a git commit to a bzr revision.
154
:return: a `bzrlib.revision.Revision` object.
157
raise AssertionError("Commit object can't be None")
158
rev = ForeignRevision(commit.id, self, self.revision_id_foreign_to_bzr(commit.id))
159
rev.parent_ids = tuple([self.revision_id_foreign_to_bzr(p) for p in commit.parents])
160
rev.message = commit.message.decode("utf-8", "replace")
161
rev.committer = str(commit.committer).decode("utf-8", "replace")
162
if commit.committer != commit.author:
163
rev.properties['author'] = str(commit.author).decode("utf-8", "replace")
165
if commit.commit_time != commit.author_time:
166
rev.properties['author-timestamp'] = str(commit.author_time)
167
if commit.commit_timezone != commit.author_timezone:
168
rev.properties['author-timezone'] = "%d" % (commit.author_timezone, )
169
rev.timestamp = commit.commit_time
170
rev.timezone = commit.commit_timezone
174
class BzrGitMappingv1(BzrGitMapping):
175
revid_prefix = 'git-v1'
179
return self.revid_prefix
182
class BzrGitMappingExperimental(BzrGitMappingv1):
183
revid_prefix = 'git-experimental'
187
class GitMappingRegistry(VcsMappingRegistry):
188
"""Registry with available git mappings."""
190
def revision_id_bzr_to_foreign(self, bzr_revid):
191
if not bzr_revid.startswith("git-"):
192
raise errors.InvalidRevisionId(bzr_revid, None)
193
(mapping_version, git_sha) = bzr_revid.split(":", 1)
194
mapping = self.get(mapping_version)
195
return mapping.revision_id_bzr_to_foreign(bzr_revid)
197
parse_revision_id = revision_id_bzr_to_foreign
200
mapping_registry = GitMappingRegistry()
201
mapping_registry.register_lazy('git-v1', "bzrlib.plugins.git.mapping",
203
mapping_registry.register_lazy('git-experimental', "bzrlib.plugins.git.mapping",
204
"BzrGitMappingExperimental")
207
class ForeignGit(ForeignVcs):
208
"""The Git Stupid Content Tracker"""
211
super(ForeignGit, self).__init__(mapping_registry)
214
def show_foreign_revid(cls, foreign_revid):
215
return { "git commit": foreign_revid }
218
foreign_git = ForeignGit()
219
default_mapping = BzrGitMappingv1()
222
def text_to_blob(texts, entry):
223
from dulwich.objects import Blob
224
text = texts.get_record_stream([(entry.file_id, entry.revision)], 'unordered', True).next().get_bytes_as('fulltext')
230
def symlink_to_blob(entry):
231
from dulwich.objects import Blob
233
blob._text = entry.symlink_target
237
def mode_is_executable(mode):
238
"""Check if mode should be considered executable."""
239
return bool(mode & 0111)
243
"""Determine the Bazaar inventory kind based on Unix file mode."""
244
entry_kind = (mode & 0700000) / 0100000
247
elif entry_kind == 1:
248
file_kind = (mode & 070000) / 010000
254
return 'tree-reference'
256
raise AssertionError(
257
"Unknown file kind %d, perms=%o." % (file_kind, mode,))
259
raise AssertionError(
260
"Unknown kind, perms=%r." % (mode,))
263
def object_mode(kind, executable):
264
if kind == 'directory':
266
elif kind == 'symlink':
269
mode = stat.S_IFREG | 0644
277
def entry_mode(entry):
278
"""Determine the git file mode for an inventory entry."""
279
return object_mode(entry.kind, entry.executable)
282
def directory_to_tree(entry, lookup_ie_sha1, unusual_modes):
283
from dulwich.objects import Tree
285
for name in sorted(entry.children.keys()):
286
ie = entry.children[name]
288
mode = unusual_modes[ie.file_id]
290
mode = entry_mode(ie)
291
hexsha = lookup_ie_sha1(ie)
292
if hexsha is not None:
293
tree.add(mode, name.encode("utf-8"), hexsha)
294
if entry.parent_id is not None and len(tree) == 0:
295
# Only the root can be an empty tree
301
def extract_unusual_modes(rev):
303
foreign_revid, mapping = mapping_registry.parse_revision_id(rev.revision_id)
304
except errors.InvalidRevisionId:
307
return mapping.export_unusual_file_modes(rev)
310
def inventory_to_tree_and_blobs(inventory, texts, mapping, unusual_modes, cur=None):
311
"""Convert a Bazaar tree to a Git tree.
313
:return: Yields tuples with object sha1, object and path
315
from dulwich.objects import Tree
322
# stack contains the set of trees that we haven't
323
# finished constructing
324
for path, entry in inventory.iter_entries():
325
while stack and not path.startswith(osutils.pathjoin(cur, "")):
326
# We've hit a file that's not a child of the previous path
329
yield sha, tree, cur.encode("utf-8")
330
mode = unusual_modes.get(cur.encode("utf-8"), stat.S_IFDIR)
331
t = (mode, urlutils.basename(cur).encode('UTF-8'), sha)
332
cur, tree = stack.pop()
335
if entry.kind == "directory":
336
stack.append((cur, tree))
340
if entry.kind == "file":
341
blob = text_to_blob(texts, entry)
342
elif entry.kind == "symlink":
343
blob = symlink_to_blob(entry)
345
raise AssertionError("Unknown kind %s" % entry.kind)
347
yield sha, blob, path.encode("utf-8")
348
name = urlutils.basename(path).encode("utf-8")
349
mode = unusual_modes.get(path.encode("utf-8"), entry_mode(entry))
350
tree.add(mode, name, sha)
352
while len(stack) > 1:
355
yield sha, tree, cur.encode("utf-8")
356
mode = unusual_modes.get(cur.encode('utf-8'), stat.S_IFDIR)
357
t = (mode, urlutils.basename(cur).encode('UTF-8'), sha)
358
cur, tree = stack.pop()
362
yield tree.id, tree, cur.encode("utf-8")
365
def revision_to_commit(rev, tree_sha, parent_lookup):
366
"""Turn a Bazaar revision in to a Git commit
368
:param tree_sha: Tree sha for the commit
369
:param parent_lookup: Function for looking up the GIT sha equiv of a bzr revision
370
:return dulwich.objects.Commit represent the revision:
372
from dulwich.objects import Commit
374
commit.tree = tree_sha
375
for p in rev.parent_ids:
376
git_p = parent_lookup(p)
377
if git_p is not None:
378
assert len(git_p) == 40, "unexpected length for %r" % git_p
379
commit.parents.append(git_p)
380
commit.message = rev.message.encode("utf-8")
381
commit.committer = fix_person_identifier(rev.committer.encode("utf-8"))
382
commit.author = fix_person_identifier(rev.get_apparent_authors()[0].encode("utf-8"))
383
commit.commit_time = long(rev.timestamp)
384
if 'author-timestamp' in rev.properties:
385
commit.author_time = long(rev.properties['author-timestamp'])
387
commit.author_time = commit.commit_time
388
commit.commit_timezone = rev.timezone
389
if 'author-timezone' in rev.properties:
390
commit.author_timezone = int(rev.properties['author-timezone'])
392
commit.author_timezone = commit.commit_timezone
added
removed
mapping.py
