1
# Copyright (C) 2009 Jelmer Vernooij <jelmer@samba.org>
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
"""Map from Git sha's to Bazaar objects."""
19
from dulwich.objects import (
24
from dulwich.object_store import (
35
from bzrlib.revision import (
39
from bzrlib.plugins.git.mapping import (
42
extract_unusual_modes,
46
from bzrlib.plugins.git.shamap import (
47
from_repository as cache_from_repository,
53
def get_object_store(repo, mapping=None):
54
git = getattr(repo, "_git", None)
56
return git.object_store
57
return BazaarObjectStore(repo, mapping)
60
MAX_TREE_CACHE_SIZE = 50 * 1024 * 1024
63
class LRUTreeCache(object):
65
def __init__(self, repository):
66
def approx_tree_size(tree):
67
# Very rough estimate, 1k per inventory entry
68
return len(tree.inventory) * 1024
69
self.repository = repository
70
self._cache = lru_cache.LRUSizeCache(max_size=MAX_TREE_CACHE_SIZE,
71
after_cleanup_size=None, compute_size=approx_tree_size)
73
def revision_tree(self, revid):
75
return self._cache[revid]
77
tree = self.repository.revision_tree(revid)
81
def iter_revision_trees(self, revids):
82
trees = dict([(k, self._cache.get(k)) for k in revids])
83
for tree in self.repository.revision_trees(
84
[r for r, v in trees.iteritems() if v is None]):
85
trees[tree.get_revision_id()] = tree
87
return (trees[r] for r in revids)
89
def revision_trees(self, revids):
90
return list(self.iter_revision_trees(revids))
93
self._cache.add(tree.get_revision_id(), tree)
96
def _find_missing_bzr_revids(get_parent_map, pending, processed):
100
processed.update(pending)
101
next_map = get_parent_map(pending)
103
for item in next_map.iteritems():
105
next_pending.update(p for p in item[1] if p not in processed)
106
pending = next_pending
107
if NULL_REVISION in todo:
108
todo.remove(NULL_REVISION)
112
def _check_expected_sha(expected_sha, object):
113
"""Check whether an object matches an expected SHA.
115
:param expected_sha: None or expected SHA as either binary or as hex digest
116
:param object: Object to verify
118
if expected_sha is None:
120
if len(expected_sha) == 40:
121
if expected_sha != object.sha().hexdigest():
122
raise AssertionError("Invalid sha for %r: %s" % (object,
124
elif len(expected_sha) == 20:
125
if expected_sha != object.sha().digest():
126
raise AssertionError("Invalid sha for %r: %s" % (object,
127
sha_to_hex(expected_sha)))
129
raise AssertionError("Unknown length %d for %r" % (len(expected_sha),
133
def _tree_to_objects(tree, parent_trees, idmap, unusual_modes):
134
"""Iterate over the objects that were introduced in a revision.
137
:param unusual_modes: Unusual file modes
138
:return: Yields (path, object, ie) entries
144
base_tree = parent_trees[0]
145
other_parent_trees = parent_trees[1:]
147
base_tree = tree._repository.revision_tree(NULL_REVISION)
148
other_parent_trees = []
149
def find_unchanged_parent_ie(ie, parent_trees):
150
assert ie.kind in ("symlink", "file")
151
for ptree in parent_trees:
153
pie = ptree.inventory[ie.file_id]
154
except errors.NoSuchId:
157
if (pie.text_sha1 == ie.text_sha1 and
158
pie.kind == ie.kind and
159
pie.symlink_target == ie.symlink_target):
162
for (file_id, path, changed_content, versioned, parent, name, kind,
163
executable) in tree.iter_changes(base_tree):
164
if kind[1] == "file":
165
ie = tree.inventory[file_id]
169
pie = find_unchanged_parent_ie(ie, other_parent_trees)
173
shamap[ie.file_id] = idmap.lookup_blob_id(
174
pie.file_id, pie.revision)
175
if not file_id in shamap:
176
new_blobs.append((path[1], ie))
177
new_trees[posixpath.dirname(path[1])] = parent[1]
178
elif kind[1] == "symlink":
179
ie = tree.inventory[file_id]
181
blob = symlink_to_blob(ie)
182
shamap[file_id] = blob.id
184
find_unchanged_parent_ie(ie, other_parent_trees)
186
yield path[1], blob, ie
187
new_trees[posixpath.dirname(path[1])] = parent[1]
188
elif kind[1] not in (None, "directory"):
189
raise AssertionError(kind[1])
190
if path[0] is not None:
191
new_trees[posixpath.dirname(path[0])] = parent[0]
193
for (path, ie), chunks in tree.iter_files_bytes(
194
[(ie.file_id, (path, ie)) for (path, ie) in new_blobs]):
198
shamap[ie.file_id] = obj.id
200
for path in unusual_modes:
201
parent_path = posixpath.dirname(path)
202
new_trees[parent_path] = tree.path2id(parent_path)
206
items = new_trees.items()
208
for path, file_id in items:
210
parent_id = tree.inventory[file_id].parent_id
211
except errors.NoSuchId:
212
# Directory was removed recursively perhaps ?
214
if parent_id is not None:
215
parent_path = urlutils.dirname(path)
216
new_trees[parent_path] = parent_id
217
trees[path] = file_id
219
def ie_to_hexsha(ie):
221
return shamap[ie.file_id]
223
# FIXME: Should be the same as in parent
224
if ie.kind in ("file", "symlink"):
226
return idmap.lookup_blob_id(ie.file_id, ie.revision)
230
blob.data = tree.get_file_text(ie.file_id)
232
elif ie.kind == "directory":
233
# Not all cache backends store the tree information,
234
# calculate again from scratch
235
ret = directory_to_tree(ie, ie_to_hexsha, unusual_modes)
242
for path in sorted(trees.keys(), reverse=True):
243
ie = tree.inventory[trees[path]]
244
assert ie.kind == "directory"
245
obj = directory_to_tree(ie, ie_to_hexsha, unusual_modes)
248
shamap[ie.file_id] = obj.id
251
class BazaarObjectStore(BaseObjectStore):
252
"""A Git-style object store backed onto a Bazaar repository."""
254
def __init__(self, repository, mapping=None):
255
self.repository = repository
257
self.mapping = default_mapping
259
self.mapping = mapping
260
self._cache = cache_from_repository(repository)
261
self._content_cache_types = ("tree")
262
self.start_write_group = self._cache.idmap.start_write_group
263
self.abort_write_group = self._cache.idmap.abort_write_group
264
self.commit_write_group = self._cache.idmap.commit_write_group
265
self.tree_cache = LRUTreeCache(self.repository)
267
def _update_sha_map(self, stop_revision=None):
268
graph = self.repository.get_graph()
269
if stop_revision is None:
270
heads = graph.heads(self.repository.all_revision_ids())
272
heads = set([stop_revision])
273
missing_revids = self._cache.idmap.missing_revisions(heads)
275
parents = graph.get_parent_map(heads)
277
for p in parents.values():
278
todo.update([x for x in p if x not in missing_revids])
279
heads = self._cache.idmap.missing_revisions(todo)
280
missing_revids.update(heads)
281
if NULL_REVISION in missing_revids:
282
missing_revids.remove(NULL_REVISION)
283
missing_revids = self.repository.has_revisions(missing_revids)
284
if not missing_revids:
286
self.start_write_group()
288
pb = ui.ui_factory.nested_progress_bar()
290
for i, revid in enumerate(graph.iter_topo_order(missing_revids)):
291
trace.mutter('processing %r', revid)
292
pb.update("updating git map", i, len(missing_revids))
293
self._update_sha_map_revision(revid)
297
self.abort_write_group()
300
self.commit_write_group()
303
self._update_sha_map()
304
return iter(self._cache.idmap.sha1s())
306
def _reconstruct_commit(self, rev, tree_sha):
307
def parent_lookup(revid):
309
return self._lookup_revision_sha1(revid)
310
except errors.NoSuchRevision:
311
trace.warning("Ignoring ghost parent %s", revid)
313
return self.mapping.export_commit(rev, tree_sha, parent_lookup)
315
def _revision_to_objects(self, rev, tree):
316
unusual_modes = extract_unusual_modes(rev)
317
present_parents = self.repository.has_revisions(rev.parent_ids)
318
parent_trees = self.tree_cache.revision_trees(
319
[p for p in rev.parent_ids if p in present_parents])
321
for path, obj, ie in _tree_to_objects(tree, parent_trees,
322
self._cache.idmap, unusual_modes):
327
# Pointless commit - get the tree sha elsewhere
328
if not rev.parent_ids:
331
base_sha1 = self._lookup_revision_sha1(rev.parent_ids[0])
332
tree_sha = self[base_sha1].tree
333
commit_obj = self._reconstruct_commit(rev, tree_sha)
335
foreign_revid, mapping = mapping_registry.parse_revision_id(
337
except errors.InvalidRevisionId:
340
_check_expected_sha(foreign_revid, commit_obj)
341
yield None, commit_obj, None
343
def _get_updater(self, rev):
344
return self._cache.get_updater(rev)
346
def _update_sha_map_revision(self, revid):
347
rev = self.repository.get_revision(revid)
348
tree = self.tree_cache.revision_tree(rev.revision_id)
349
updater = self._get_updater(rev)
350
for path, obj, ie in self._revision_to_objects(rev, tree):
351
updater.add_object(obj, ie)
352
commit_obj = updater.finish()
355
def _reconstruct_blobs(self, keys):
356
"""Return a Git Blob object from a fileid and revision stored in bzr.
358
:param fileid: File id of the text
359
:param revision: Revision of the text
361
stream = self.repository.iter_files_bytes(
362
((key[0], key[1], key) for key in keys))
363
for (fileid, revision, expected_sha), chunks in stream:
365
blob.chunked = chunks
366
if blob.id != expected_sha and blob.data == "":
367
# Perhaps it's a symlink ?
368
tree = self.tree_cache.revision_tree(revision)
369
entry = tree.inventory[fileid]
370
if entry.kind == 'symlink':
371
blob = symlink_to_blob(entry)
372
_check_expected_sha(expected_sha, blob)
375
def _reconstruct_tree(self, fileid, revid, inv, unusual_modes,
377
"""Return a Git Tree object from a file id and a revision stored in bzr.
379
:param fileid: fileid in the tree.
380
:param revision: Revision of the tree.
382
def get_ie_sha1(entry):
383
if entry.kind == "directory":
385
return self._cache.idmap.lookup_tree_id(entry.file_id,
387
except (NotImplementedError, KeyError):
388
obj = self._reconstruct_tree(entry.file_id, revid, inv,
394
elif entry.kind in ("file", "symlink"):
396
return self._cache.idmap.lookup_blob_id(entry.file_id,
400
return self._reconstruct_blobs(
401
[(entry.file_id, entry.revision, None)]).next().id
403
raise AssertionError("unknown entry kind '%s'" % entry.kind)
404
tree = directory_to_tree(inv[fileid], get_ie_sha1, unusual_modes)
405
_check_expected_sha(expected_sha, tree)
408
def get_parents(self, sha):
409
"""Retrieve the parents of a Git commit by SHA1.
411
:param sha: SHA1 of the commit
412
:raises: KeyError, NotCommitError
414
return self[sha].parents
416
def _lookup_revision_sha1(self, revid):
417
"""Return the SHA1 matching a Bazaar revision."""
418
from dulwich.protocol import ZERO_SHA
419
if revid == NULL_REVISION:
422
return self._cache.idmap.lookup_commit(revid)
425
return mapping_registry.parse_revision_id(revid)[0]
426
except errors.InvalidRevisionId:
427
self._update_sha_map(revid)
428
return self._cache.idmap.lookup_commit(revid)
430
def get_raw(self, sha):
431
"""Get the raw representation of a Git object by SHA1.
433
:param sha: SHA1 of the git object
436
return (obj.type, obj.as_raw_string())
438
def __contains__(self, sha):
439
# See if sha is in map
441
(type, type_data) = self.lookup_git_sha(sha)
443
return self.repository.has_revision(type_data[0])
445
return self.repository.texts.has_version(type_data)
447
return self.repository.has_revision(type_data[1])
449
raise AssertionError("Unknown object type '%s'" % type)
453
def lookup_git_shas(self, shas, update_map=True):
457
ret[sha] = self._cache.idmap.lookup_git_sha(sha)
460
# if not, see if there are any unconverted revisions and add
461
# them to the map, search for sha in map again
462
self._update_sha_map()
465
ret[sha] = self._cache.idmap.lookup_git_sha(sha)
470
def lookup_git_sha(self, sha, update_map=True):
471
return self.lookup_git_shas([sha], update_map=update_map)[sha]
473
def __getitem__(self, sha):
474
if self._cache.content_cache is not None:
476
return self._cache.content_cache[sha]
479
(type, type_data) = self.lookup_git_sha(sha)
480
# convert object to git object
482
(revid, tree_sha) = type_data
484
rev = self.repository.get_revision(revid)
485
except errors.NoSuchRevision:
486
trace.mutter('entry for %s %s in shamap: %r, but not found in '
487
'repository', type, sha, type_data)
489
commit = self._reconstruct_commit(rev, tree_sha)
490
_check_expected_sha(sha, commit)
493
(fileid, revision) = type_data
494
return self._reconstruct_blobs([(fileid, revision, sha)]).next()
496
(fileid, revid) = type_data
498
tree = self.tree_cache.revision_tree(revid)
499
rev = self.repository.get_revision(revid)
500
except errors.NoSuchRevision:
501
trace.mutter('entry for %s %s in shamap: %r, but not found in repository', type, sha, type_data)
503
unusual_modes = extract_unusual_modes(rev)
505
return self._reconstruct_tree(fileid, revid, tree.inventory,
506
unusual_modes, expected_sha=sha)
507
except errors.NoSuchRevision:
510
raise AssertionError("Unknown object type '%s'" % type)
512
def generate_pack_contents(self, have, want, progress=None,
514
"""Iterate over the contents of a pack file.
516
:param have: List of SHA1s of objects that should not be sent
517
:param want: List of SHA1s of objects that should be sent
520
ret = self.lookup_git_shas(have + want)
521
for commit_sha in have:
523
(type, (revid, tree_sha)) = ret[commit_sha]
527
assert type == "commit"
530
for commit_sha in want:
531
if commit_sha in have:
534
(type, (revid, tree_sha)) = ret[commit_sha]
538
assert type == "commit"
541
todo = _find_missing_bzr_revids(self.repository.get_parent_map,
543
trace.mutter('sending revisions %r', todo)
545
pb = ui.ui_factory.nested_progress_bar()
547
for i, revid in enumerate(todo):
548
pb.update("generating git objects", i, len(todo))
549
rev = self.repository.get_revision(revid)
550
tree = self.tree_cache.revision_tree(revid)
551
for path, obj, ie in self._revision_to_objects(rev, tree):
552
ret.append((obj, path))
557
def add_thin_pack(self):
560
fd, path = tempfile.mkstemp(suffix=".pack")
561
f = os.fdopen(fd, 'wb')
563
from dulwich.pack import PackData, Pack
564
from bzrlib.plugins.git.fetch import import_git_objects
567
if os.path.getsize(path) == 0:
570
pd.create_index_v2(path[:-5]+".idx", self.object_store.get_raw)
573
self.repository.lock_write()
575
self.repository.start_write_group()
577
import_git_objects(self.repository, self.mapping,
578
p.iterobjects(get_raw=self.get_raw),
581
self.repository.abort_write_group()
584
self.repository.commit_write_group()
586
self.repository.unlock()
589
# The pack isn't kept around anyway, so no point
590
# in treating full packs different from thin packs
591
add_pack = add_thin_pack