55
72
return bzr_rev_id[len(cls.revid_prefix)+1:], cls()
57
74
def generate_file_id(self, path):
75
# Git paths are just bytestrings
76
# We must just hope they are valid UTF-8..
77
assert isinstance(path, str)
60
return escape_file_id(path.encode('utf-8'))
80
return escape_file_id(path)
82
def parse_file_id(self, file_id):
83
if file_id == ROOT_ID:
85
return unescape_file_id(file_id)
62
87
def import_commit(self, commit):
63
88
"""Convert a git commit to a bzr revision.
68
93
raise AssertionError("Commit object can't be None")
69
94
rev = ForeignRevision(commit.id, self, self.revision_id_foreign_to_bzr(commit.id))
70
95
rev.parent_ids = tuple([self.revision_id_foreign_to_bzr(p) for p in commit.parents])
71
rev.message = commit.message.decode("utf-8", "replace")
72
rev.committer = str(commit.committer).decode("utf-8", "replace")
96
rev.message = escape_invalid_chars(commit.message.decode("utf-8", "replace"))[0]
97
rev.committer = escape_invalid_chars(str(commit.committer).decode("utf-8", "replace"))[0]
73
98
if commit.committer != commit.author:
74
rev.properties['author'] = str(commit.author).decode("utf-8", "replace")
99
rev.properties['author'] = escape_invalid_chars(str(commit.author).decode("utf-8", "replace"))[0]
101
if commit.commit_time != commit.author_time:
102
rev.properties['author-timestamp'] = str(commit.author_time)
103
if commit.commit_timezone != commit.author_timezone:
104
rev.properties['author-timezone'] = str(commit.author_timezone)
75
105
rev.timestamp = commit.commit_time
106
rev.timezone = commit.commit_timezone
121
151
default_mapping = BzrGitMappingv1()
124
def inventory_to_tree_and_blobs(repo, mapping, revision_id):
125
from dulwich.objects import Tree, Blob
126
from bzrlib.inventory import InventoryDirectory, InventoryFile
154
def text_to_blob(texts, entry):
155
from dulwich.objects import Blob
156
text = texts.get_record_stream([(entry.file_id, entry.revision)], 'unordered', True).next().get_bytes_as('fulltext')
162
def symlink_to_blob(entry):
163
from dulwich.objects import Blob
165
blob._text = entry.symlink_target
169
def entry_mode(entry):
170
if entry.kind == 'directory':
172
elif entry.kind == 'symlink':
174
elif entry.kind == 'file':
175
mode = stat.S_IFREG | 0644
183
def directory_to_tree(entry, lookup_ie_sha1):
184
from dulwich.objects import Tree
186
for name in sorted(entry.children.keys()):
187
ie = entry.children[name]
188
tree.add(entry_mode(ie), name.encode("utf-8"), lookup_ie_sha1(ie))
193
def inventory_to_tree_and_blobs(inventory, texts, mapping, cur=None):
194
"""Convert a Bazaar tree to a Git tree.
196
:return: Yields tuples with object sha1, object and path
198
from dulwich.objects import Tree
132
inv = repo.get_inventory(revision_id)
134
205
# stack contains the set of trees that we haven't
135
206
# finished constructing
137
for path, entry in inv.iter_entries():
138
while stack and not path.startswith(cur):
207
for path, entry in inventory.iter_entries():
208
while stack and not path.startswith(osutils.pathjoin(cur, "")):
209
# We've hit a file that's not a child of the previous path
140
sha = tree.sha().hexdigest()
212
yield sha, tree, cur.encode("utf-8")
142
213
t = (stat.S_IFDIR, urlutils.basename(cur).encode('UTF-8'), sha)
143
214
cur, tree = stack.pop()
146
if type(entry) == InventoryDirectory:
217
if entry.kind == "directory":
147
218
stack.append((cur, tree))
151
if type(entry) == InventoryFile:
152
#FIXME: We can make potentially make this Lazy to avoid shaing lots of stuff
153
# and having all these objects in memory at once
155
_, blob._text = repo.iter_files_bytes([(entry.file_id, entry.revision, path)]).next()
156
sha = blob.sha().hexdigest()
157
yield sha, blob, path
222
if entry.kind == "file":
223
blob = text_to_blob(texts, entry)
224
elif entry.kind == "symlink":
225
blob = symlink_to_blob(entry)
227
raise AssertionError("Unknown kind %s" % entry.kind)
229
yield sha, blob, path.encode("utf-8")
159
230
name = urlutils.basename(path).encode("utf-8")
160
mode = stat.S_IFREG | 0644
163
tree.add(mode, name, sha)
231
tree.add(entry_mode(entry), name, sha)
165
233
while len(stack) > 1:
167
sha = tree.sha().hexdigest()
236
yield sha, tree, cur.encode("utf-8")
169
237
t = (stat.S_IFDIR, urlutils.basename(cur).encode('UTF-8'), sha)
170
238
cur, tree = stack.pop()
174
yield tree.sha().hexdigest(), tree, cur
242
yield tree.id, tree, cur.encode("utf-8")
177
245
def revision_to_commit(rev, tree_sha, parent_lookup):
187
255
for p in rev.parent_ids:
188
256
git_p = parent_lookup(p)
189
257
if git_p is not None:
258
assert len(git_p) == 40, "unexpected length for %r" % git_p
190
259
commit._parents.append(git_p)
191
260
commit._message = rev.message.encode("utf-8")
192
261
commit._committer = rev.committer.encode("utf-8")
193
commit._author = rev.get_apparent_author().encode("utf-8")
262
commit._author = rev.get_apparent_authors()[0].encode("utf-8")
194
263
commit._commit_time = long(rev.timestamp)
264
if 'author-timestamp' in rev.properties:
265
commit._author_time = long(rev.properties['author-timestamp'])
267
commit._author_time = commit._commit_time
268
commit._commit_timezone = rev.timezone
269
if 'author-timezone' in rev.properties:
270
commit._author_timezone = int(rev.properties['author-timezone'])
272
commit._author_timezone = commit._commit_timezone
195
273
commit.serialize()