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,
51
def get_object_store(repo, mapping=None):
52
git = getattr(repo, "_git", None)
54
return git.object_store
55
return BazaarObjectStore(repo, mapping)
58
MAX_TREE_CACHE_SIZE = 50 * 1024 * 1024
61
class LRUTreeCache(object):
63
def __init__(self, repository):
64
def approx_tree_size(tree):
65
# Very rough estimate, 1k per inventory entry
66
return len(tree.inventory) * 1024
67
self.repository = repository
68
self._cache = lru_cache.LRUSizeCache(max_size=MAX_TREE_CACHE_SIZE,
69
after_cleanup_size=None, compute_size=approx_tree_size)
71
def revision_tree(self, revid):
73
return self._cache[revid]
75
tree = self.repository.revision_tree(revid)
79
def iter_revision_trees(self, revids):
80
trees = dict([(k, self._cache.get(k)) for k in revids])
81
for tree in self.repository.revision_trees(
82
[r for r, v in trees.iteritems() if v is None]):
83
trees[tree.get_revision_id()] = tree
85
return (trees[r] for r in revids)
87
def revision_trees(self, revids):
88
return list(self.iter_revision_trees(revids))
91
self._cache.add(tree.get_revision_id(), tree)
94
def _check_expected_sha(expected_sha, object):
95
"""Check whether an object matches an expected SHA.
97
:param expected_sha: None or expected SHA as either binary or as hex digest
98
:param object: Object to verify
100
if expected_sha is None:
102
if len(expected_sha) == 40:
103
if expected_sha != object.sha().hexdigest():
104
raise AssertionError("Invalid sha for %r: %s" % (object,
106
elif len(expected_sha) == 20:
107
if expected_sha != object.sha().digest():
108
raise AssertionError("Invalid sha for %r: %s" % (object,
109
sha_to_hex(expected_sha)))
111
raise AssertionError("Unknown length %d for %r" % (len(expected_sha),
115
def _tree_to_objects(tree, parent_trees, idmap, unusual_modes):
116
"""Iterate over the objects that were introduced in a revision.
119
:param unusual_modes: Unusual file modes
120
:return: Yields (path, object, ie) entries
126
base_tree = parent_trees[0]
127
other_parent_trees = parent_trees[1:]
129
base_tree = tree._repository.revision_tree(NULL_REVISION)
130
other_parent_trees = []
131
def find_unchanged_parent_ie(ie, parent_trees):
132
assert ie.kind in ("symlink", "file")
133
for ptree in parent_trees:
135
pie = ptree.inventory[ie.file_id]
136
except errors.NoSuchId:
139
if (pie.text_sha1 == ie.text_sha1 and
140
pie.kind == ie.kind and
141
pie.symlink_target == ie.symlink_target):
144
for (file_id, path, changed_content, versioned, parent, name, kind,
145
executable) in tree.iter_changes(base_tree):
146
if kind[1] == "file":
147
ie = tree.inventory[file_id]
151
pie = find_unchanged_parent_ie(ie, other_parent_trees)
155
shamap[ie.file_id] = idmap.lookup_blob_id(
156
pie.file_id, pie.revision)
157
if not file_id in shamap:
158
new_blobs.append((path[1], ie))
159
new_trees[urlutils.dirname(path[1])] = parent[1]
160
elif kind[1] == "symlink":
161
ie = tree.inventory[file_id]
163
blob = symlink_to_blob(ie)
164
shamap[file_id] = blob.id
166
find_unchanged_parent_ie(ie, other_parent_trees)
168
yield path[1], blob, ie
169
new_trees[urlutils.dirname(path[1])] = parent[1]
170
elif kind[1] not in (None, "directory"):
171
raise AssertionError(kind[1])
172
if path[0] is not None:
173
new_trees[urlutils.dirname(path[0])] = parent[0]
175
for (path, ie), chunks in tree.iter_files_bytes(
176
[(ie.file_id, (path, ie)) for (path, ie) in new_blobs]):
180
shamap[ie.file_id] = obj.id
182
for fid in unusual_modes:
183
new_trees[tree.id2path(fid)] = tree.inventory[fid].parent_id
187
items = new_trees.items()
189
for path, file_id in items:
191
parent_id = tree.inventory[file_id].parent_id
192
except errors.NoSuchId:
193
# Directory was removed recursively perhaps ?
195
if parent_id is not None:
196
parent_path = urlutils.dirname(path)
197
new_trees[parent_path] = parent_id
198
trees[path] = file_id
200
def ie_to_hexsha(ie):
202
return shamap[ie.file_id]
204
if ie.kind in ("file", "symlink"):
206
return idmap.lookup_blob_id(ie.file_id, ie.revision)
210
blob.data = tree.get_file_text(ie.file_id)
212
elif ie.kind == "directory":
213
# Not all cache backends store the tree information,
214
# calculate again from scratch
215
ret = directory_to_tree(ie, ie_to_hexsha, unusual_modes)
222
for path in sorted(trees.keys(), reverse=True):
223
ie = tree.inventory[trees[path]]
224
assert ie.kind == "directory"
225
obj = directory_to_tree(ie, ie_to_hexsha, unusual_modes)
228
shamap[ie.file_id] = obj.id
231
class BazaarObjectStore(BaseObjectStore):
232
"""A Git-style object store backed onto a Bazaar repository."""
234
def __init__(self, repository, mapping=None):
235
self.repository = repository
237
self.mapping = default_mapping
239
self.mapping = mapping
240
self._cache = cache_from_repository(repository)
241
self._content_cache_types = ("tree")
242
self.start_write_group = self._cache.idmap.start_write_group
243
self.abort_write_group = self._cache.idmap.abort_write_group
244
self.commit_write_group = self._cache.idmap.commit_write_group
245
self.tree_cache = LRUTreeCache(self.repository)
247
def _update_sha_map(self, stop_revision=None):
248
graph = self.repository.get_graph()
249
if stop_revision is None:
250
heads = graph.heads(self.repository.all_revision_ids())
252
heads = set([stop_revision])
253
missing_revids = self._cache.idmap.missing_revisions(heads)
255
parents = graph.get_parent_map(heads)
257
for p in parents.values():
258
todo.update([x for x in p if x not in missing_revids])
259
heads = self._cache.idmap.missing_revisions(todo)
260
missing_revids.update(heads)
261
if NULL_REVISION in missing_revids:
262
missing_revids.remove(NULL_REVISION)
263
missing_revids = self.repository.has_revisions(missing_revids)
264
if not missing_revids:
266
self.start_write_group()
268
pb = ui.ui_factory.nested_progress_bar()
270
for i, revid in enumerate(graph.iter_topo_order(missing_revids)):
271
trace.mutter('processing %r', revid)
272
pb.update("updating git map", i, len(missing_revids))
273
self._update_sha_map_revision(revid)
277
self.abort_write_group()
280
self.commit_write_group()
283
self._update_sha_map()
284
return iter(self._cache.idmap.sha1s())
286
def _reconstruct_commit(self, rev, tree_sha):
287
def parent_lookup(revid):
289
return self._lookup_revision_sha1(revid)
290
except errors.NoSuchRevision:
291
trace.warning("Ignoring ghost parent %s", revid)
293
return self.mapping.export_commit(rev, tree_sha, parent_lookup)
295
def _revision_to_objects(self, rev, tree):
296
unusual_modes = extract_unusual_modes(rev)
297
present_parents = self.repository.has_revisions(rev.parent_ids)
298
parent_trees = self.tree_cache.revision_trees(
299
[p for p in rev.parent_ids if p in present_parents])
301
for path, obj, ie in _tree_to_objects(tree, parent_trees,
302
self._cache.idmap, unusual_modes):
307
# Pointless commit - get the tree sha elsewhere
308
if not rev.parent_ids:
311
base_sha1 = self._lookup_revision_sha1(rev.parent_ids[0])
312
tree_sha = self[base_sha1].tree
313
commit_obj = self._reconstruct_commit(rev, tree_sha)
315
foreign_revid, mapping = mapping_registry.parse_revision_id(
317
except errors.InvalidRevisionId:
320
_check_expected_sha(foreign_revid, commit_obj)
321
yield None, commit_obj, None
323
def _get_updater(self, rev):
324
return self._cache.get_updater(rev)
326
def _update_sha_map_revision(self, revid):
327
rev = self.repository.get_revision(revid)
328
tree = self.tree_cache.revision_tree(rev.revision_id)
329
updater = self._get_updater(rev)
330
for path, obj, ie in self._revision_to_objects(rev, tree):
331
updater.add_object(obj, ie)
332
commit_obj = updater.finish()
335
def _reconstruct_blobs(self, keys):
336
"""Return a Git Blob object from a fileid and revision stored in bzr.
338
:param fileid: File id of the text
339
:param revision: Revision of the text
341
stream = self.repository.iter_files_bytes(
342
((key[0], key[1], key) for key in keys))
343
for (fileid, revision, expected_sha), chunks in stream:
345
blob.chunked = chunks
346
if blob.id != expected_sha and blob.data == "":
347
# Perhaps it's a symlink ?
348
tree = self.tree_cache.revision_tree(revision)
349
entry = tree.inventory[fileid]
350
if entry.kind == 'symlink':
351
blob = symlink_to_blob(entry)
352
_check_expected_sha(expected_sha, blob)
355
def _reconstruct_tree(self, fileid, revid, inv, unusual_modes,
357
"""Return a Git Tree object from a file id and a revision stored in bzr.
359
:param fileid: fileid in the tree.
360
:param revision: Revision of the tree.
362
def get_ie_sha1(entry):
363
if entry.kind == "directory":
365
return self._cache.idmap.lookup_tree_id(entry.file_id,
367
except (NotImplementedError, KeyError):
368
obj = self._reconstruct_tree(entry.file_id, revid, inv,
374
elif entry.kind in ("file", "symlink"):
376
return self._cache.idmap.lookup_blob_id(entry.file_id,
380
return self._reconstruct_blobs(
381
[(entry.file_id, entry.revision, None)]).next().id
383
raise AssertionError("unknown entry kind '%s'" % entry.kind)
384
tree = directory_to_tree(inv[fileid], get_ie_sha1, unusual_modes)
385
_check_expected_sha(expected_sha, tree)
388
def get_parents(self, sha):
389
"""Retrieve the parents of a Git commit by SHA1.
391
:param sha: SHA1 of the commit
392
:raises: KeyError, NotCommitError
394
return self[sha].parents
396
def _lookup_revision_sha1(self, revid):
397
"""Return the SHA1 matching a Bazaar revision."""
398
if revid == NULL_REVISION:
401
return self._cache.idmap.lookup_commit(revid)
404
return mapping_registry.parse_revision_id(revid)[0]
405
except errors.InvalidRevisionId:
406
self._update_sha_map(revid)
407
return self._cache.idmap.lookup_commit(revid)
409
def get_raw(self, sha):
410
"""Get the raw representation of a Git object by SHA1.
412
:param sha: SHA1 of the git object
415
return (obj.type, obj.as_raw_string())
417
def __contains__(self, sha):
418
# See if sha is in map
420
(type, type_data) = self._lookup_git_sha(sha)
422
return self.repository.has_revision(type_data[0])
424
return self.repository.texts.has_version(type_data)
426
return self.repository.has_revision(type_data[1])
428
raise AssertionError("Unknown object type '%s'" % type)
432
def _lookup_git_sha(self, sha):
433
# See if sha is in map
435
return self._cache.idmap.lookup_git_sha(sha)
437
# if not, see if there are any unconverted revisions and add them
438
# to the map, search for sha in map again
439
self._update_sha_map()
440
return self._cache.idmap.lookup_git_sha(sha)
442
def __getitem__(self, sha):
443
if self._cache.content_cache is not None:
445
return self._cache.content_cache[sha]
448
(type, type_data) = self._lookup_git_sha(sha)
449
# convert object to git object
451
(revid, tree_sha) = type_data
453
rev = self.repository.get_revision(revid)
454
except errors.NoSuchRevision:
455
trace.mutter('entry for %s %s in shamap: %r, but not found in '
456
'repository', type, sha, type_data)
458
commit = self._reconstruct_commit(rev, tree_sha)
459
_check_expected_sha(sha, commit)
462
(fileid, revision) = type_data
463
return self._reconstruct_blobs([(fileid, revision, sha)]).next()
465
(fileid, revid) = type_data
467
tree = self.tree_cache.revision_tree(revid)
468
rev = self.repository.get_revision(revid)
469
except errors.NoSuchRevision:
470
trace.mutter('entry for %s %s in shamap: %r, but not found in repository', type, sha, type_data)
472
unusual_modes = extract_unusual_modes(rev)
474
return self._reconstruct_tree(fileid, revid, tree.inventory,
475
unusual_modes, expected_sha=sha)
476
except errors.NoSuchRevision:
479
raise AssertionError("Unknown object type '%s'" % type)
481
def generate_pack_contents(self, have, want, progress=None, get_tagged=None):
482
"""Iterate over the contents of a pack file.
484
:param have: List of SHA1s of objects that should not be sent
485
:param want: List of SHA1s of objects that should be sent
488
for commit_sha in have:
490
(type, (revid, tree_sha)) = self._lookup_git_sha(commit_sha)
494
assert type == "commit"
497
for commit_sha in want:
498
if commit_sha in have:
500
(type, (revid, tree_sha)) = self._lookup_git_sha(commit_sha)
501
assert type == "commit"
505
processed.update(pending)
506
next_map = self.repository.get_parent_map(pending)
508
for item in next_map.iteritems():
510
next_pending.update(p for p in item[1] if p not in processed)
511
pending = next_pending
512
if NULL_REVISION in todo:
513
todo.remove(NULL_REVISION)
514
trace.mutter('sending revisions %r', todo)
516
pb = ui.ui_factory.nested_progress_bar()
518
for i, revid in enumerate(todo):
519
pb.update("generating git objects", i, len(todo))
520
rev = self.repository.get_revision(revid)
521
tree = self.tree_cache.revision_tree(revid)
522
for path, obj, ie in self._revision_to_objects(rev, tree):
523
ret.append((obj, path))
528
def add_thin_pack(self):
531
fd, path = tempfile.mkstemp(suffix=".pack")
532
f = os.fdopen(fd, 'wb')
534
from dulwich.pack import PackData, Pack
535
from bzrlib.plugins.git.fetch import import_git_objects
538
if os.path.getsize(path) == 0:
541
pd.create_index_v2(path[:-5]+".idx", self.object_store.get_raw)
544
self.repository.lock_write()
546
self.repository.start_write_group()
548
import_git_objects(self.repository, self.mapping,
549
p.iterobjects(get_raw=self.get_raw),
552
self.repository.abort_write_group()
555
self.repository.commit_write_group()
557
self.repository.unlock()
560
# The pack isn't kept around anyway, so no point
561
# in treating full packs different from thin packs
562
add_pack = add_thin_pack
added
removed
object_store.py
